AI needs to be smaller, reduce energy footprint: Study
A combination of shorter queries and using more specific models and could cut AI energy consumption by up to 90 pe rcent without sacrificing performance, said UNESCO in a report published to mark the AI for Good global summit in Geneva.
OpenAI CEO Sam Altman recently revealed that each request sent to its popular generative AI app ChatGPT consumes on average 0.34 Wh of electricity, which is between 10 and 70 times a Google search.
With ChatGPT receiving around a billion requests per day that amounts to 310 GWh annually, equivalent to the annual electricity consumption of three million people in Ethiopia, for example,
Moreover, UNESCO calculated that AI energy demand is doubling every 100 days as generative AI tools become embedded in everyday life.
"The exponential growth in computational power needed to run these models is placing increasing strain on global energy systems, water resources, and critical minerals, raising concerns about environmental sustainability, equitable access, and competition over limited resources," the UNESCO report warned.
However, it was able to achieve a nearly 90 per cent reduction in electricity usage by reducing the length of its query, or prompt, as well as by using a smaller AI, without a drop in performance.
Many AI models like ChatGPT are general-purpose models designed to respond on a wide variety of topics, meaning that it must sift through an immense volume of information to formulate and evaluate responses.
The use of smaller, specialised AI models offers major reductions in electricity needed to produce a response.
So did cutting the cutting prompts from 300 to 150 words.
Being already aware of the energy issue, tech giants all now offer miniature versions with fewer parameters of their respective large language models.
For example, Google sells Gemma, Microsoft has Phi-3, and OpenAI has GPT-4o mini. French AI companies have done likewise, for instance, Mistral AI has introduced its model Ministral.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
CNA
an hour ago
- CNA
Nvidia's $4 trillion milestone caps rise of stock market behemoth
NEW YORK :The stunning rise of Nvidia Corp to become the first publicly traded company valued at $4 trillion underscores the massive importance to the stock market of the AI chipmaker and the technology sector. Reaching the lofty milestone caps a massive run for Nvidia's stock price, which has soared some 1,350 per cent since October 2022. This year, the chipmaker's shares are up about 22 per cent against a 6 per cent rise for the S&P 500. The company's market value hit $4 trillion in morning trading on Wednesday, about 13 months after it had first reached the $3 trillion milestone. Nvidia's stock surge has given it significant weight in key equity indexes and ETFs that are more heavily influenced by companies with the biggest market values. At around 7.5 per cent, Nvidia has the biggest weight in the S&P 500, which is widely viewed by investors as the benchmark for the U.S. stock market. The stock has an even greater presence in more tech-heavy measures, such as the popular Invesco QQQ Trust ETF and the Philadelphia SE Semiconductor Index. Its influence is less significant in the Dow Jones Industrial Average, which is weighted by stock prices as opposed to market values. Nvidia may soon have company in the exclusive $4 trillion club. Microsoft's market value was hovering around $3.7 trillion on Wednesday. Apple ranks third at about $3.1 trillion. Their massive market values also illustrate the hefty influence of tech and tech-related companies. The top seven S&P 500 weights, which also include Alphabet, Meta Platforms and Broadcom, comprise about one-third of the index. Nvidia's gains also point to the increasing presence of the overall technology sector, which is by far the largest sector in the S&P 500. The tech sector's market value has grown to one-third of S&P 500 market value, nearly reaching the proportion that the sector hit during the height of the dot com bubble in 2000. Other tech stocks that have performed well in 2025 include Microsoft, up about 19 per cent this year, Oracle, up 40 per cent, and Palantir, up 88 per cent.
Straits Times
2 hours ago
- Straits Times
X removes posts by Musk's AI chatbot Grok after anti-Semitism complaints
Sign up now: Get ST's newsletters delivered to your inbox Billionaire Elon Musk's chatbot Grok shocked users of Mr Musk's social media platform X on July 8 when it started producing content with anti-Semitic tropes and praise for Adolf Hitler. WASHINGTON - Social media posts on the X account of the Grok chatbot developed by Elon Musk's company xAI were removed on July 8 after complaints from X users and the Anti-Defamation League that Grok produced content with anti-Semitic tropes and praise for Adolf Hitler. Issues of political biases, hate speech and accuracy of AI chatbots have been a concern since at least the launch of OpenAI's ChatGPT in 2022. 'We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts,' according to a post on Grok's account on X, which Mr Musk owns. 'Since being made aware of the content, xAI has taken action to ban hate speech before Grok posts on X. xAI is training only truth-seeking and thanks to the millions of users on X, we are able to quickly identify and update the model where training could be improved.' On July 8, posts on Grok's X account suggested Hitler would be best-placed to combat anti-white hatred, saying he would 'spot the pattern and handle it decisively.' The posts also referred to Hitler positively as 'history's mustache man,' and commented that people with Jewish surnames were responsible for extreme anti-white activism. ADL, the non-profit organisation formed to combat anti-Semitism, urged xAI and other producers of Large Language Model software that produces human-sounding text to avoid 'producing content rooted in anti-Semitic and extremist hate.' Top stories Swipe. Select. Stay informed. World Trump's ambassador nominee to Singapore Anjani Sinha has a rough day at Senate hearing Singapore Singapore to hire more than 1,000 new educators annually in the next few years, up from 700 Singapore COE prices rise for all categories Singapore Govt watching job situation for fresh graduates closely, exploring further support: Gan Siow Huang Singapore Man who killed cats by throwing them off HDB blocks has jail term doubled to 27 months Singapore $43k fine for undischarged bankrupt doctor who failed to disclose assets worth over $4m Singapore Female primary school teacher charged over alleged sex acts with underage male student Singapore People working in air-conditioned spaces prefer 24 deg C and warmer: Survey 'What we are seeing from Grok LLM right now is irresponsible, dangerous and anti-Semitic, plain and simple. This supercharging of extremist rhetoric will only amplify and encourage the anti-Semitism that is already surging on X and many other platforms,' ADL said on X. In May, after users noticed Grok brought up the topic of 'white genocide' in South Africa in unrelated discussions about other matters, xAI attributed it to an unauthorised change that was made to Grok's response software. Mr Musk in June promised an upgrade to the company's model, suggesting there was, 'far too much garbage in any foundation model trained on uncorrected data.' During the latest controversy, posts on Grok's X account said the model made a 'slip-up' by engaging with comments posted by a fake account with a common Jewish surname, and later that the fake account was a 'troll hoax to fuel division.' REUTERS
CNA
2 hours ago
- CNA
French competition authority to investigate Meta's online advertising practices
PARIS :The French competition authority said on Wednesday it had informed Meta Platforms that the Facebook owner was accused of abusing its dominant position in the online advertising sector. "(...) An objection has been notified to (...) Meta," the authority said in a statement, adding this procedure, likely to last several months, adding that investigations were ongoing and that the company has not been found guilty. A spokesperson for the competition authority told Reuters that the procedure has been triggered by a complaint filed in October 2022 of Adloox, an advertising platform, which has since been bought by private U.S. company Scope3. In April, French media companies, including TF1, France TV and BFM TV, said they were suing Meta over what they say are unlawful business practices. The company will also face trial this October in Spain over a 551 million euro ($645.06 million) complaint lodged by more than 80 media companies accusing it of unfair competition in advertising. In February online, rights campaigners filed complaints in Europe over Meta's advertising practices, while earlier on Wednesday, EU antitrust regulators fined Meta, as well as Apple for what they said were breaches of EU law.
