Latest news with #AaronMak
Politico
2 days ago
- Business
- Politico
The tech woes of America's tech cops
With help from Alfred Ng and Aaron Mak Frustration is brewing around the government office whose job is to track the most exquisite, cutting-edge American technology as it moves around the world. Though a wonky agency known mostly to trade and tech insiders, the Bureau of Industry and Security at the Commerce Department operates at the cutting edge of the $600 billion global semiconductor industry. It polices American export controls by approving or rejecting export licenses and pursuing enforcement actions. You might think the BIS would have sophisticated tools to deploy — but in fact, America's global tech cop is hobbled by software that in some cases hasn't been updated in two decades. In a series of hearings and interviews, insiders have described an agency struggling to cope with creaking software even as its workload has ballooned in recent years. Last year, then-undersecretary of Commerce Alan Estevez, who led BIS, told Congress licensing applications had doubled to 40,000 since 2012 while some of its 'antiquated' software dated back to 2006. The state of affairs has shocked tech policy experts. 'The IT infrastructure at the Bureau of Industry and Security is a legitimate national security threat to the U.S.,' said a senior Republican aide on the House Foreign Affairs Committee, which oversees BIS. (The aide requested anonymity to speak about ongoing discussions in Congress.) A Democratic House staffer said BIS under the Biden administration responded at a glacial pace to requests for information, including important questions on applications for export licenses to Chinese tech companies. 'It took them over a year to get back to Congress with licensing data about SMIC and Huawei,' said the aide, who was not authorized to speak to the press. The BIS did not respond to DFD's questions on its software. Its current leader, Commerce Department undersecretary Jeffrey Kessler, told lawmakers in June that the president's budget requested $303 million for the agency, 'the single largest investment in BIS history, and one that will significantly bolster our ability to protect national security.' The House and Senate are also making efforts to shore up the agency. The House has suggested appropriating $303 million for BIS, while the Senate has suggested $211 million. Either would be an increase from the current $191 million. But it's not clear if that would cover the full software overhaul insiders say is needed — and Congress might not manage to pass appropriations bills before an Oct. 1 shutdown, which could mean flat funding for the near term. The BIS has become increasingly important in American policy over its four decades of existence, as the U.S. has expanded its use of export controls for foreign policy. During President Donald Trump's first term, he placed Huawei and SMIC on the 'Entity List' of blacklisted companies. His administration used the Foreign Direct Product Rule to extend U.S. export controls to goods made with U.S. components, even if the products were manufactured outside the U.S. Both of those are policed by the BIS. Biden built on those controls, both by enacting new restrictions on the export of advanced AI chips, and by adding Russian companies to the 'Entity List' over Moscow's invasion of Ukraine. As BIS was handling a bigger slice of the global tech trade, its systems struggled to keep up. Speaking to HFAC last year, Estevez said his agency was using 'manual pulls' of data for Congress. The Democratic House aide said BIS brought staffers in for a demonstration. 'It was very distressing,' the aide said. 'Mostly I felt really really bad for the people who work at Commerce.' The staffer pointed to frustrations at all levels, from disconnected systems handling major BIS responsibilities to character limits that made it hard to search accurately for company names. 'The search functions are alarmingly inadequate,' the staffer said. In an effort to modernize, the agency awarded Palantir $3.5 million in 2024 for data center services. However, that was far from the scale of overhaul a database would require. Even if lawmakers increase BIS's budget, the ancient tech is only one part of its problem. The agency staff has been in turmoil since Trump took office, with top leaders on chip export controls departing. David Feith, who handled technology at the National Security Council, was dismissed in April. A bipartisan bill HFAC lawmakers introduced last week would require the bureau to nearly double the ranks of BIS export control officers around the world to 20, from 11 today. Kessler told lawmakers last month that extra BIS money would enable the agency to hire nearly 200 special agents across the U.S. as well as 30 officers overseas. Michael Sobolik, a senior fellow at the Hudson Institute who previously worked as an aide to Sen. Ted Cruz (R-Texas), said ramping up BIS staff could come at the expense of a tech reboot. 'If that money is going to have to go to cover new hires, and if the tech is really as old as it seems like it is, that may not be enough,' said Sobolik. Jessie Blaeser contributed to this report. AI pricing argument hits the airlines Some lawmakers want to know how Delta Airlines is using artificial intelligence to set 'individualized fares' for customers — an emerging industry practice that critics fear abuses people's privacy and creates higher costs. The letter, sent by Sen. Ruben Gallego (D-Ariz.) and joined by Sen. Richard Blumenthal (D-Conn.) and Mark Warner (D-Va.) on Monday and exclusively provided to POLITICO, asks Delta CEO Ed Bastian to disclose what data the company uses to train its algorithm for setting prices, how many customers are paying for prices determined by AI and what steps the company is taking to evaluate the potential impact of AI on its pricing. Unlike discounts for broad groups such as teachers, veterans and senior citizens, individualized pricing sets specific prices for different people, with critics saying the goal is to set the highest price that people are willing to pay. It also uses more personal information, such as a person's web browsing behavior, social media activity, financial status and location, the Federal Trade Commission said in a January report. Delta did not immediately respond to requests for comment. Delta told investors on a July 10 earnings call that it plans to use AI to set pricing for 20 percent of its domestic flights by the end of 2025. The company expects to eventually personalize all flight prices. The practice of using people's data to set individualized prices, which the FTC calls 'surveillance pricing,' has been a concern among state lawmakers who worry it will lead to higher prices. Business lobbyists have pushed back against regulations, arguing that individual pricing can bring discounts to consumers. Delta has until Aug. 4 to respond to the senators' letter. Incoming: Trump's AI Action Plan Trump is set to release his long-awaited AI Action Plan on Wednesday — and POLITICO's Mohar Chatterjee reports that it will center on cutting regulation, competing with China and suppressing 'woke AI.' The president called on his AI and crypto czar, David Sacks, and other White House tech advisers to develop the action plan in a January executive order. The order rolled back AI safety guardrails that President Joe Biden put into place through his own EO. Mohar obtained an overview of the action plan on Friday, and reports that it will include provisions for streamlining permitting for data center construction, ensuring that federal AI systems are not ideologically biased, promoting AI exports and withholding funds from states with restrictive AI laws. Many of these aims align with the administration's light-touch approach to AI — for example, it backed a state AI law moratorium in the spending bill. That provision was cut, and members of Congress are now trying to get bipartisan support for AI regulations in statehouses. Yet, as POLITICO's Morning Tech team points out, using this action plan to push for state AI deregulation could be a potential workaround to Congress' rejection of the moratorium. post of the day THE FUTURE IN 5 LINKS Stay in touch with the whole team: Aaron Mak (amak@ Mohar Chatterjee (mchatterjee@ Steve Heuser (sheuser@ Nate Robson (nrobson@ and Daniella Cheslow (dcheslow@
Politico
14-07-2025
- Business
- Politico
How war, and Silicon Valley, are driving an Israeli tech rebound
With help from Aaron Mak It wasn't long ago that Israel's tech sector saw a drop in both confidence and global investments. After the October 7, 2023 Hamas attack and the ensuing Israeli retaliation in Gaza, foreign direct investment plummeted, and the government injected millions of dollars into the tech industry to try to lure investors. But now, a rebound is firmly underway — led in many cases by American investment. The California semiconductor giant Nvidia is planning a new campus of up to 30 acres in the country's north, it told DFD. Google acquired Israeli-founded cybersecurity company Wiz for $32 billion in March. Since the war began, venture capital firms Sequoia and Greylock announced they were hiring new local partners; Andreessen Horowitz also reportedly made significant new investments in Israel (the company didn't respond to DFD questions). Startup Nation Central, a nonprofit promoting Israeli innovation, found in January that Israeli tech companies raised more than $9.3 billion in the first half of 2025, which it said was the strongest six-month performance in the last three years. Investors in Israel say the recovery is fueled by growing global interest in cybersecurity and artificial intelligence, both of which are Israeli strengths, and have now been tested on the battlefield in Gaza. Politically, too, Israel is benefiting from more vocal support in Silicon Valley, despite global criticism of Israel's retaliatory actions. Most recently, Google co-founder Sergey Brin called the United Nations 'transparently antisemitic' after it published a report on companies, including Google, that profit from working with Israel, The Washington Post reported. (Google didn't respond to a question about Brin's comment.) The Financial Times marked the shift in an essay published Monday: 'Despite all the international criticism of Israel for its multiple military offensives, from Gaza to Iran, a surge in foreign buying has fuelled the rally in its stock market…. Israel is cementing its status as the region's dominant economic force.' Shuly Galili, managing partner and founder of the Silicon Valley-based venture capital firm UpWest VC, which invests in Israel, said she had noticed the bounceback as well. 'Between October [2023] and January – that was probably the most disruptive period,' she said. She mused that foreign investors might be swooping in to be 'contrarian' and 'to announce this is time to get best deals and put in money on ground.' Israel is not a consensus cause in the tech world, where the rank-and-file politics often lean progressive. Last year, dozens of Google workers protested the company's contract with the Israeli government. The company fired them. Critics continue to target Israeli policies — and its use of technology — as morally questionable. Former Prime Minister Ehud Olmert has accused Israel of war crimes in the course of an onslaught that has killed more than 58,000 Palestinians, according to local health officials. And Palestinians say Israeli's dominance in AI should not be a selling point - it should be a cautionary tale. One group called it 'the most nefarious threat to Palestinians' living under Israeli control because of its speed, potential inaccuracies and the shift of responsibility to machines. But these concerns do not seem to have dented Israel's appeal for investors — especially those looking for advanced military technology powered by AI. It's all a remarkable pivot from the early days after Hamas attacked Israel and killed 1,200 people, taking hundreds more hostage. After the attack — and Israel's overwhelming military reaction — airlines canceled their flights, major tech companies nixed events, and the sector's workers were in and out of reserves. By 2024, one financial newspaper reported tech deals were at a five-year low. But since then, economically at least, the conflict has gone from a liability to something more like a proof of concept. In one literal example, the state-owned Rafael Advanced Defense Systems appeared to show a real-life drone attack on a man in a promotional video for its Spike Firefly loitering munition, which one analyst geolocated to northern Gaza. The company did not respond to a request for comment. Startup Nation Central's CEO Avi Hasson said in a statement that defense companies have shifted into 'high gear: battlefield urgency has fast-tracked development, surfaced exceptional talent, and inspired a new wave of entrepreneurs. Startups are pouring into this space, and capital is following.' Grant Demaree, a U.S. military veteran and founder of the Pentagon contractor Onebrief, says Israel had offered some real-world lessons in how to integrate AI on the battlefield. 'The biggest lesson is that for AI to be useful, it need not to be perfect, it just needs to significantly outperform people,' he said. As an example, he mentioned AI for targeting. 'We know humans make a lot of targeting mistakes all the time,' he said. 'If you can have an AI that makes fewer mistakes and is significantly faster, I think that contributes to a better outcome and is just better than just doing the process manually, even if that is still imperfect.' Palestinians say AI — far from 'outperforming' — is responsible for some of the worst carnage of the conflict. Hani Almadhoun, who lives in the Washington, D.C. area and runs a soup kitchen in his native Gaza using GoFundMe, said his brother Majed was killed with his wife and children in an Israeli airstrike in November 2023; a second brother Mahmoud was killed a year later by a drone. He told DFD he suspected Majed might have been mistakenly targeted by AI, though he had no way to confirm the feeling. 'AI hasn't made war smarter. It's just made it way easier to kill people without feeling guilty about it,' he said. 'That's not progress; that's just pure horror, all wrapped up in fancy computer code,' he said. The Israeli military did not immediately respond to POLITICO's questions. In response to previous reports that soldiers were relying on an AI-based program to identify targets, the Israeli military said that AI was only used in 'auxiliary tools that assist officers in the process of incrimination.' Meanwhile, Silicon Valley's politics are shifting, with Israel fitting into a new, Trump-aligned sensibility. After Brin complained about a UN report accusing 48 companies – including Google, Microsoft, Palantir and Amazon – of complicity in Israel's 'economy of genocide,' Secretary of State Marco Rubio sanctioned its author. In Washington, after a Palestinian heckler accused Palantir of killing Palestinians with AI, CEO Alex Karp retorted, 'Mostly terrorists, that's true.' Galili, the Israeli investor, said the war disrupted travel for investors and founders. But she said, 'overall we've seen a lot of support from tech leaders. I'm based in Silicon Valley and my day to day is interacting with investors in Silicon Valley. There was day-to-day a lot of support.' There are still other factors that could slow down Israel's tech dominance. President Donald Trump has announced huge deals to sell semiconductors and build data centers in Saudi Arabia and the United Arab Emirates, which could eventually become tech powerhouses. A more tangible cloud on the horizon might be internal politics. Before October 7, Israel was riven by protests over a plan by Prime Minister Benjamin Netanyahu to overhaul the judicial system. That began scaring investors away. A new survey by AllJobs finds nearly three-quarters of Israeli workers considering relocation overseas. 'Many things can spark political unrest,' Galili said. 'External threats we can deal with, internal threats – we're still working on it.' California AI bills get delayed After hitting legislative hurdles, lawmakers are putting off their push for two pathbreaking AI bills in Sacramento, POLITICO's California Decoded team reports. One bill, sponsored by assemblymember Rebecca Bauer-Kahan, has pitted Hollywood against the tech industry. AB 412 would allow copyright owners to request information on the datasets that developers use to train their AI models. Hollywood power players, like the actors union SAG-AFTRA, have backed the bill as a tool for artists to protect their rights. However, the AI industry has warned that AB 412 would stifle innovation. Bauer-Kahan has paused her efforts to pass the bill until 2026. Her spokesperson told POLITICO's Tyler Katzenberger that this will give them time to 'resolve outstanding issues.' Democratic state Sen. Jerry McNerney is also delaying action on SB 813, which would create expert panels to vet AI systems, after the bill got stuck in the Senate Appropriations Committee. If passed, developers would have the option to voluntarily submit their models for inspection in exchange for legal protections in personal injury cases. The Decoded team reported that a mysterious nonprofit called Fathom has been advocating for the bill since its inception. Fathom has $25 million in funding, though it hasn't disclosed who exactly its donors are. It's 'crypto week' at the House The House is set to vote on two cryptocurrency bills during what Republicans lawmakers are calling 'crypto week,' staffers told POLITICO's Jasper Goodman. On Wednesday, the House is expected to vote on legislation that would effectively ban Federal Reserve banks from issuing digital currencies. Republicans have expressed concerns that such assets would give the government too much power in the cryptocurrency space. However, Rep. Maxine Waters (D-Calif.) wrote in a May House Financial Services Committee report that cryptocurrency backed by central banks could have 'a greater potential to gain broad public trust and utilization (especially after numerous digital assets have been used to defraud Americans).' The House will then vote on Thursday on the Senate's GENIUS Act, which establishes a regulatory framework for stablecoins — cryptocurrency that is pegged to another kind of asset, like the dollar. It includes provisions to split up oversight duties among market regulators and implement measures against money laundering. If passed, the legislation could lend more legitimacy to stablecoins, and allow financial institutions to get more involved in the cryptocurrency sector. Trump's sons notably founded a cryptocurrency firm last year known as World Liberty Financial Inc., which issues a stablecoin. post of the day THE FUTURE IN 5 LINKS Stay in touch with the whole team: Aaron Mak (amak@ Mohar Chatterjee (mchatterjee@ Steve Heuser (sheuser@ Nate Robson (nrobson@ and Daniella Cheslow (dcheslow@
Politico
10-07-2025
- Business
- Politico
The tech winner in SCOTUS ruling on porn age checks
With help from Daniella Cheslow and Aaron Mak The Supreme Court's decision to uphold Texas' age verification law for porn websites will likely have a ripple effect across the internet — and not just for the vast numbers of people who watch online porn. The Texas law says that websites that have more than one-third of their content considered 'sexual material harmful to minors' need to verify ages before granting access, ensuring that any visitors are older than 18. When it upheld the law in late June, the Supreme Court also cleared the path for the 23 other states that have similar regulations in place, and 21 more states that are considering passing their own requirements. A majority of Americans, then, will soon need to share some kind of information — such as a face scan, a photo ID or a phone number — to access pornography online. The adult entertainment industry challenged the law and lost, as did free speech advocates like the American Civil Liberties Union, who worry that mandatory age verification erodes people's online privacy and anonymity. Tech industry groups opted to stay out of the case, though they have fought similar battles against age verification laws for social media websites. One group that warmly welcomed the decision, however, was the emerging crop of tech companies that provide age verification services. 'We expect this judgment to accelerate the adoption of age verification laws across the remaining 26 US states and globally,' Iain Corby, the executive director of the Age Verification Providers Association said in a statement following the SCOTUS decision. The trade group, which started in 2018, represents a new industry providing a service that more and more laws are making mandatory. Companies offer age verification technology through several methods, including using AI to estimate a person's age based on their voice, their face and their digital history. The Supreme Court's decision is expected to lead to a boom for the industry, both in demand and innovation, observers say. 'We're going to see what is technically possible in terms of age assurance,' Ariel Fox Johnson, a senior advisor for data privacy at Common Sense Media, a children's advocacy group, said at a Cato Institute event on July 8. 'We're going to see what companies are really capable of here.' Privacy advocates worry that the age verification tools, whether it's using an AI to scan your face to estimate your age or submitting your personal information to analyze your online activity, are an increasingly invasive technology. They also worry the rise of age verification will further entrench the need for data brokers — another industry that has raised privacy concerns as it has grown. Depending on the technique, age verification can rely heavily on material provided by data brokers. The industry already plays a major role in identity verification, which could easily be transferred to age verification tools. In a brief it submitted to the Supreme Court, the AVPA highlighted a company that uses a person's phone number to check against 'commercially available databases' to verify their age. Another uses an email address to analyze online interactions and transaction data, looking for records like mortgages or credit cards that would determine they are over 18. Data brokers, or companies that collect and sell people's information for a variety of purposes, have become essential for advertisers, fraud prevention and the federal government, often highlighting the benefits of being able to quickly deliver details on millions of Americans. While lawmakers and privacy advocates have raised concerns about how information sold by data brokers creates risks, both to public officials and to national security, the industry's increasingly central role in commerce and government business has made it tough for advocates to push back. (Lawmakers drew one line last year, passing a law making it illegal for data brokers to sell data to foreign adversaries.) Now, with a huge new landscape of age-verification requirements, that entrenchment is likely to deepen. 'There are legitimate deep, and complicated conversations about the harms that data brokers pose, what are the appropriate ways to regulate them, how does that square with constitutional considerations — but these age verification laws along with some other laws are simply getting out ahead of that legislative discussion, and enshrining these practices as legitimate child safety efforts,' Cody Venzke, a senior policy counsel at the American Civil Liberties Union, told Digital Future Daily. A fuzzy audit trail for AI-generated police reports AI-generated police reports were designed to be difficult to audit, according to a new report from the Electronic Frontier Foundation. The report published Thursday is based on a review of public records obtained by the digital rights group from police departments subscribed to Axon's Draft One software, as well as public remarks from the company. In a roundtable discussion about Draft One from last year, Axon's senior principal product manager for generative AI Noah Spitzer-Williams explained that its software intentionally deletes the first draft of its AI-written police reports, to avoid potential transparency requests. 'So we don't store the original draft and that's by design. And that's really because the last thing we want to do is create more disclosure headaches for our customers and our attorney's offices,' he said in the video. EFF said that AI-generated police reports should be aggressively tracked and audited, but Axon's design intentionally prevents that. An Axon spokesperson said that Draft One was designed to replicate the existing process for filing police reports, where 'only the final, approved report is saved and discoverable, not the interim edits, additions or deletions made during officer or supervisor review.' EFF's report comes after Alfred reported for DFD last September that several police departments were struggling to parse out which reports were written by real people and which were AI-generated in response to public records requests. The group obtained internal discussions showing the Frederick police department in Colorado asking for help to figure out which reports were written by AI, and Axon responding by sharing its own internal tracking data, after noting that the audit log available for its customers may have issues. 'To set expectaitons (sic), it's not going to be graceful, but this wasn't a scenario we anticipated needing to make easy,' Spitzer-Williams told the Frederick police department last August, according to a copy of the email exchange obtained by the EFF and provided to POLITICO. Europe unveils rules for advanced AI The European Union published its highly anticipated guidance for advanced AI models like Gemini, Grok and ChatGPT on Thursday. As POLITICO's Pieter Haeck reports, AI companies that agree to the voluntary code of practice will be required to report certain design details of their models and their policies surrounding copyrighted content. Models that could impact public health and safety are subject to additional systemic risk mitigation measures. Companies that sign on to the guidance will have a clearer route to comply with the EU's AI Act, a wide-ranging statute that covers everything from facial recognition systems to image generators. Some of the act's most broadly applicable provisions — particularly regarding large language models — are set to go into effect in August. Big tech corporations and others have been raising a fuss over the AI Act. Last week, both the European and American tech industries sent letters in a bid to modify the law. Leaders from European companies like Mistral, ASML and Mercedes-Benz pleaded in a letter to the European Commission for a two-year delay on implementation to allow regulators to further simplify the rules. U.S. industry groups representing Meta, Google and others also sent a letter to President Donald Trump, urging him to press the EU on the 'discriminatory' impacts of tech laws like the AI Act in trade negotiations. Kai Zenner, digital advisor to European Parliament member Axel Voss of Germany, told POLITICO's Daniella Cheslow that the guidance 'isn't looking that bad.' He said civil society groups and lawmakers pushed the European commission to resist pressure from tech companies, but that those advocates are likely to have more fights on their hands in the coming years. post of the day THE FUTURE IN 5 LINKS Stay in touch with the whole team: Aaron Mak (amak@ Mohar Chatterjee (mchatterjee@ Steve Heuser (sheuser@ Nate Robson (nrobson@ and Daniella Cheslow (dcheslow@
Politico
07-07-2025
- Business
- Politico
Big Tech looks for some global wins
With help from Aaron Mak Big Tech might have lost its bid to freeze state AI rules, but on the global front, it's finding new opportunities to push back on regulation it doesn't like. Nearly six months after President Donald Trump took office promising to go after 'unfair' overseas digital regulation, the American tech industry has already notched one victory and is pushing for more. Canada folded first: Last week, after Trump threatened to stop trade with Ottawa, Canada agreed to shelve a digital services tax that would have collected billions of dollars from U.S. tech companies. Now, looking to press its case in Europe, half a dozen U.S. tech trade groups signed onto a letter last week urging the White House to take a stand against 'the EU's discriminatory digital regulatory policy.' Europe's Digital Services Act, which aims to curb harmful content online, has already come up as a topic of conversation between American and European negotiators, POLITICO's Ari Hawkins reported. European officials pushed back on reports that the Digital Markets Act, which regulates competition, is on the table, too. No trade deal is set yet, and Trump on Monday announced the tariffs that had been set to snap back this week will once again be postponed as he negotiates deals with U.S. trading partners including Europe. But the environment is giving American tech advocates the sense that the wind is at their backs. 'It's hard for me to see the Europeans just giving it up entirely, but they're going to have to make changes' to their digital rules, said Ed Brzytwa, vice president of international trade at the Consumer Technology Association, which counts Amazon, Meta and Apple as members. His group signed last week's letter to the White House, which named the DSA, DMA, digital services taxes, EU AI Act and the General Data Protection Regulation among 'discriminatory, unreasonable and burdensome regulations and taxes in the EU.' American companies say they're unfairly targeted by European policies. In the latest example, Apple, which faces fines of €500 million, and Meta, which was fined €200 million, are challenging their penalties for breaching the EU's Digital Markets Act. The trade war that Trump opened seems to have created new avenues to push their case, and now the tech industry has other digital measures in its sights. 'Canada walking back the DST was an example of how everything is on the table in these trade negotiations,' said Daniel Castro, vice president at the Information Technology and Innovation Foundation, a nonprofit whose supporters include the CTA, Alphabet, Amazon, Apple, Microsoft and Meta. To some extent, European tech groups are joining the push to defang the digital rules coming out of Brussels. More than 40 leaders of Europe's biggest corporations, including French AI powerhouse Mistral and the Dutch chipmaking equipment manufacturer ASML, urged Brussels in a letter last week to pause the EU AI Act for two years. They pointed to the challenges pointed out by Mario Draghi, the former European Central Bank chief, whose seminal report last September found Europe lagging in productivity. He blamed Europe's financial markets, costly energy, fragmented R&D and polarized politics. The push against the AI Act, from both sides of the Atlantic, seemed to land with a receptive audience. European Commission tech chief Henna Virkkunen told Politico last month she would not rule out postponing some of the AI Act. Nick Moës, executive director of The Future Society that favors tech regulation, said the push was 'emblematic of significant lobbying going on to try to water down or delete obligations and requirements imposed upon industry players developing AI.' But other tech regulations may be harder to crack. In late June, The Wall Street Journal reported the U.S. Trade Representative's office had compiled a draft trade agreement with the EU that included a list of tentative deals with Brussels, including a dialogue on the Digital Markets Act. According to the report, American companies would be exempt from DMA enforcement while talks went on. A senior member of the European Parliament told POLITICO's Jacob Parry the WSJ report 'is bullshit' and that 'DMA is not up for discussion.' Brzytwa, of the CTA, said he could see policymakers in Brussels struggling to outright repeal tech regulation that had been carefully crafted over years – but the enforcement might be lightened. 'What we're hopeful for is it becomes less burdensome on companies with respect to fines,' he said. Brzytwa said he saw the White House taking an expansive view of what might be possible on the global front. In Trump's first term, he ordered investigations into individual policies that burdened American companies. Now, however, Trump was targeting a wide range of digital policies. 'His aperture in terms of the issue set has broadened,' Brzytwa said. 'We haven't seen anything in writing but all the tea leaves seem to indicate…the issues we're citing are on the table.' Communications systems blamed for flood disaster As deadly floods devastate Texas, alleged failures in emergency communications infrastructure may have been a driving force of the destruction. Over the weekend, flash floods coursed through central Texas, resulting in the deaths of at least 80 people. Dozens of people are also missing. Some local officials have accused the National Weather Service (NWS) of making faulty forecasts, which the agency has disputed. Yet, emergency preparedness experts say that gaps in the technical measures to disseminate warnings are really to blame. 'The crux of this disaster is a failure of the last mile of communication,' Tom Fahy, legislative director for the NWS employee union, told POLITICO. Such alerts were transmitted wirelessly to handheld devices, and to the National Oceanic and Atmospheric Administration's Weather Radio. However, the messages can't reach people who don't have access to a cellphone or reliable service, or who've disabled such notifications on their devices. A summer camp where 11 people are missing notably had a policy against touchscreen devices that could've received weather alerts. In addition, people may miss radio communications if they don't happen to be listening at the right moment. Other counties in Central Texas have tried to solve this problem by installing outdoor weather sirens. post of the day THE FUTURE IN 5 LINKS Stay in touch with the whole team: Aaron Mak (amak@ Mohar Chatterjee (mchatterjee@ Steve Heuser (sheuser@ Nate Robson (nrobson@ and Daniella Cheslow (dcheslow@
Politico
01-07-2025
- Politico
AI is coming for our wireless networks
With help from Aaron Mak Artificial intelligence feels like it's everywhere these days, and the wireless networks connecting our devices will soon be no exception. Telecom wonks are determining how AI should be built into sprawling cellular networks to better transmit phone calls and allocate resources for the hundreds of data-hungry apps that people use each day. Every decade or so, a new generation of wireless technology arrives for our phones, marked in the transitions from 2G to 3G and so on. We're on 5G now, which brought the fastest speeds people have seen to date, starting in 2019. That means it's time to prepare for 6G in the next five years. That's the era when many people believe AI will play a significant role in juggling and allocating the resources powering devices and AI-related tools. 'The transition to 6G will be a lot more visible,' Ronnie Vasishta, senior vice president of telecom at Nvidia, told DFD during a recent visit to Washington. 'For a long time, 5G was looking for that monetizable killer app, and I think that check mark comes with these generative AI use models.' 6G's more expansive promises, where cellular networks converge with AI, quantum and other tech efforts premised on explosive speeds and better computing power, have taken shape since the start of the decade and finally feel at hand. Initial 6G deployments could come as soon as 2028 or so, Vasishta believes, with many companies eyeing 2030 as the big 6G year. The new era could see smart glasses and autonomous vehicles competing for 6G bandwidth in a sea of internet-connected devices and more widespread integration of generative AI. At play will also be better sensing technology, which will have both commercial and defense applications, and better tools to fight spam calls (although spammers will likely have their own enhanced tools). This parade of new applications will strain the capacity of 5G networks. 'We're going to see the network holding us back,' Vasishta added. 'All of a sudden, we'll start seeing a pressing need for 6G. But 6G will be kind of a different type of network.' And it's meant a broader industry conversation about how to handle this transformation. The fact that Nvidia, a tech giant best known for chipmaking, cares at all about 6G is a sign of how broad the transition — and AI's role in it — is likely to be. Vasishta is focused on the company's AI platform that's intended to blend wireless network operations with faster computing power. The goal is what the industry calls 'AI-native 6G wireless,' which could bring a host of benefits, including the more efficient use of spectrum and better network operations. Nvidia's efforts picked up last fall as the company announced a partnership with cellular heavyweight T-Mobile and expanded it in March to a broader coalition of like-minded players, including the MITRE Corp, Cisco, Open RAN Development Company and Booz Allen Hamilton, all trying to figure out how these new systems are supposed to work. The participants also know they need to keep 6G secure, and that includes the global push against Chinese telecom giants like Huawei. Nvidia's 6G ambitions are also far-reaching. It just announced a collaboration with the British government on the topic and touts ongoing research in Finland, Germany and France. Much of the global story around 5G involved fighting between the U.S. and China over the sway of Huawei and fellow Chinese powerhouse ZTE. The U.S. has decreed both companies national security threats and sought to minimize their global presence — but China is very much still a part of the story. '6G is not far away — the standards process has started and is being fully engaged by the global U.S. competitors,' Douglas Robbins, vice president of engineering and prototyping at MITRE Labs, told DFD. 'Huawei is very active, just presented recently their vision for AI in 6G.' That means, per Robbins, 'the time is now' for the U.S. to assert a leadership role in crafting 6G, which means creating the technical standards for it, and fostering research for the transition. He hopes to see flexible standards that allow companies to regularly update their software-based wireless architecture and that embrace post-quantum cryptography and security protocols. Washington, of course, is paying attention, at least in fits and starts. The Biden administration had hammered out high-level 6G principles. Last month the House Energy and Commerce Committee held a subcommittee hearing on AI's role in communications infrastructure, where Vasishta testified about Nvidia's endeavors. And the GOP is now seeking to free huge tranches of wireless frequencies for the private sector in the party's reconciliation package — 'serious power for American Leadership on 6G,' President Donald Trump proclaimed on Truth Social in June. But the road to wireless supremacy may come with bumps. U.S. policy, as this year shows, can swerve in unexpected directions. Although lawmakers devoted $1.5 billion to advanced wireless research in the 2022 CHIPS and Science Act — grants aimed at fostering 5G open radio access networks, intended to help boost U.S. competition with China — Republicans are now looking to scrap the remaining funds in their reconciliation bill. That cut was included in the Senate legislation that passed Tuesday. The country's R&D efforts also face heightened pressure as the Trump administration eyes restricting foreign student visas and scaling back federal support for research. 'We need government help in fostering research,' Vasishta said. 'In the U.S, there's some amazing research that's happening in the area of AI native wireless. … At the end of the day, those researchers also need research funding, and there's some fundamental research that's coming out of the U.S. that I think could make a big difference to standards committees.' The iPhone antitrust case moves forward A federal judge has ruled Apple must face claims that it has an illegal smartphone monopoly. In his ruling Monday, Judge Julien Neals of the U.S. District Court for New Jersey wrote that the Justice Department and several states adequately presented evidence showing Apple allegedly has a smartphone monopoly given its 65 percent share of the market. He also found it was plausible that Apple intended to maintain its dominance through monopolistic practices. The decision points to barriers such as third-party phones having difficulties connecting to Apple Watch, accessing iMessage features, and making digital payments without Apple Wallet. Apple, which is represented by a legal team from Kirkland & Ellis LLP and Gibson, Dunn & Crutcher LLP, said in a statement, 'We believe this lawsuit is wrong on the facts and the law, and we will continue to vigorously fight it in court.' Denmark wants age verification across the EU Denmark is looking to make age verification for children on social media a priority as it takes over the presidency of the Council of the European Union starting Tuesday. 'All European countries have the opportunity to demand from the Big Tech platforms that they introduce this [age verification],' Caroline Stage Olsen, Denmark's digital affairs minister, told POLITICO Europe's Eliza Gkritsi. Olsen has also been pushing to adjust the EU's General Data Protection Regulation to raise the threshold for parental consent for data processing of minors from its current standard of children under 13 up to 15. These efforts build on momentum in Denmark and other European countries to more tightly regulate kids' online activities. Danish Prime Minister Mette Frederiksen called for a ban last year on social media use by kids under 15. Danish officials have also said they believe that online platforms are fueling a spate of youth gang violence. French President Emmanuel Macron threatened last month to ban social media for those younger than 15 if the EU did not institute more youth internet restrictions for the bloc. Spain, Greece and the United Kingdom have also moved to institute additional age-based restrictions for platform use. post of the day THE FUTURE IN 5 LINKS Stay in touch with the whole team: Aaron Mak (amak@ Mohar Chatterjee (mchatterjee@ Steve Heuser (sheuser@ Nate Robson (nrobson@ and Daniella Cheslow (dcheslow@
