ChatGPT use may weaken critical thinking and memory, suggests new MIT study
A recent study by researchers at the MIT Media Lab has raised concerns over the impact of ChatGPT on young people's cognitive engagement and learning. Conducted over several months, the study involved 54 participants aged 18 to 39 from the Boston area. Each was asked to write SAT-style essays using either ChatGPT, Google Search, or no digital tool at all.
The findings were striking. EEG scans tracking brain activity across 32 regions showed that participants using ChatGPT demonstrated the lowest levels of brain engagement. These users 'consistently underperformed at neural, linguistic, and behavioural levels,' according to the researchers. Over time, the ChatGPT group grew increasingly passive, often resorting to copy-paste methods by the end of the study.
Also read: Nvidia employee gives Indian parents tour of US office, internet moved by photo
Lead researcher Nataliya Kosmyna explained why she chose to publish the findings before peer review. 'What really motivated me to put it out now before waiting for a full peer review is that I am afraid in 6–8 months, there will be some policymaker who decides, 'let's do GPT kindergarten.' I think that would be absolutely bad and detrimental,' she said. 'Developing brains are at the highest risk.'
Participants in the 'brain-only' group, those who wrote essays without any assistance, showed the highest neural connectivity, particularly in the alpha, theta, and delta bands associated with creativity, memory, and semantic processing. This group also expressed more curiosity, ownership, and satisfaction with their work.
Those using Google Search also exhibited high engagement and satisfaction. According to the researchers, this difference is notable as more users now turn to AI chatbots instead of traditional search engines to find information.
After completing the essays, participants were asked to revise one of their earlier drafts. Those in the ChatGPT group, now writing without the tool, struggled to recall their previous work and showed lower brainwave activity. 'The task was executed, and you could say that it was efficient and convenient,' Kosmyna said. 'But as we show in the paper, you basically didn't integrate any of it into your memory networks.'
In contrast, the brain-only group, when given access to ChatGPT for the rewrite, demonstrated increased brain connectivity, suggesting that AI can enhance learning if used after active, independent thinking.
While the study has not yet been peer reviewed and involved a relatively small sample, it adds to a growing body of research at MIT examining the broader impact of generative AI. Previous studies from the lab have also linked extended AI use with increased feelings of loneliness.
Also read: 'Complete breakdown during video call': Bengaluru man hospitalised after CEO's brutal outburst
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
an hour ago
- Time of India
‘Reliance Jio may become dominant FWA provider by June-end, ahead of T-Mobile'
NEW DELHI: Reliance Jio is expected to become a dominant fixed wireless access (FWA) service provider by subscribers globally by the end of June, ahead of rival Bharti Airtel and US telecom carrier T-Mobile , according to ICICI Securities. The Mukesh Ambani-led telco gained 1.03 million 5G FWA customers in June, taking its total base to 5.85 million. By comparison, Airtel gained 182,458 customers, taking its user base to 1.54 million, according to the industry data collated by the Telecom Regulatory Authority of India (TRAI) for May. The sector regulator has reclassified users of FWA-UBR (Fixed Wireless Access Unlicensed Band Radio) as fixed wireline customers. 'Industry FWA (excluding UBR) stood at 7.4 million with RJio's FWA subs at 5.9 million, and adjusted for UBR reclassification, net add was an impressive 0.74 million in May 2025. RJio's FWA (including UBR) stood at 6.88 million. This, compared to T-Mobile (US player, and had largest FWA subs base globally), FWA subs base was at 6.85 million in March 2025 (quarter),' the brokerage said in a research note on Sunday, which ETTelecom has reviewed. 'We believe RJio is path to become dominant player by subs for FWA globally by end-June 2025. Bharti's FWA subs rose 0.18 million to 1.5 million in May 2025; and Bharti is also scaling its FWA services to monetise the rising demand for fixed broadband in India,' it added. Jio and Airtel have seen a rapid uptake of their respective 5G fixed broadband services since the commercial launch in 2023. 5G FWA is also seen as a significant revenue-generation opportunity for the telcos, given the low penetration of wired broadband in India and a rise in data consumption. ICICI estimated that the wired broadband (including FWA) subsribers base rose at an average of 1.8 million per month over December 2024 to May 2025 to 51.5 million, representing a 24.6% year-on-year growth. 'Bharti's subs increase 0.29mn/month to 10.8mn. RJio added 0.66mn subs/month, now at 19.4mn. Other operators' subs base has been largely stable,' it said. According to ICICI, Jio's wired broadband (including FWA) market share rose to 37.6% in May 2025, compared to 34% in November 2024. During the same period, Airtel's share rose from about 20% to 21%, while state-controlled Bharat Sanchar Nigam Limited's (BSNL) user base remained the same at 4.3 million but its market share fell from 10.3% in November 2024 to 8.4% in May 2025. Swedish telecom gear maker Ericsson , in its recently released report, said that a strong need for accessible broadband in rural and semi-urban areas is driving Indian operators to expand their 5G FWA footprints. 'Availability of affordable 5G FWA customer premises equipment (CPE) is also driving growth of 5G FWA, which will help bridge the digital divide,' it has said. Jio widens gap with rivals in active user base In May, Jio's active subscribers expanded by 5.5 million to 462 million, whereas Airtel's active subscribers base increased 1.3 million to 387 million, while Vodafone Idea's active subscribers declined by 1.3 million to 173 million as its network expansion is yet to show a visible impact, as per ICICI. Jio's active user market share rose 16bps (basis points) month-on-month to 42.8%, while Airtel's declined 13bps to 35.8% and Vi's fell 23bps to 16%, it said. 1bps is 0.01%. Industry-wide active subscriber base was up by 7.4 million to 1.08 billion, following a 0.7% month-on-month growth in May.
Time of India
an hour ago
- Time of India
Reliance Jio on path to become dominant Fixed Wireless Access player globally: Analysts
Overall, Reliance Jio dominated the India market with 50.72 per cent share in the broadband segment, comprising both wired and wireless segments. Reliance Jio is set to become the world's largest Fixed Wireless Access provider. It will surpass T-Mobile in subscriber base. Jio's 5G FWA subscribers reached 6.88 million in May. The company dominates the Indian market with over 50 percent broadband share. Reliance Jio and Bharti Airtel added most new customers in May. Tired of too many ads? Remove Ads Tired of too many ads? Remove Ads New Delhi: Indian telecom major Reliance Jio is on track to become the world's largest Fixed Wireless Access provider by subscriber base, eclipsing US-based telecom giant T-Mobile , an analyst report said on to recent data released by telecom regulator Trai, Reliance Jio total 5G FWA subscriber base (including unlicensed band radio) reached 6.88 million in May, while T-Mobile had 6.85 million subscribers in Jio reclassified around 1 million Fixed Wireless Access (FWA) unlicensed band radio (UBR) subscribers into FTTx category, following which its net FWA subscriber base stood at 5.9 million in May, with a monthly addition of 0.74 million Securities report by analysts Sanjesh Jain, Mohit Mishra and Aparajita Chakraborty based on Trai data said that industry FWA (excluding UBR) stood at 7.4 million with RJio's FWA subs stood at 5.9 million after UBR report said that net addition was 0.74 million subscribers in May."RJio's FWA (including UBR) stood at 6.88 mn. This, compared to T-Mobile (US player, and the largest FWA subs base globally) FWA subs base was at 6.85 mn in March 2025. We believe RJio is on a path to become the dominant player by subs for FWA globally by end-June 2025," the report Reliance Jio dominated the India market with 50.72 per cent share in the broadband segment, comprising both wired and wireless Jio total fixed wired broadband subscriber base was 13.51 million and wireless broadband subscriber base was 480.96 million in May, as per data published by the Telecom Regulatory Authority of India (Trai).Reliance Jio is followed by Bharti Airtel in India with 30.99 per cent share in the broadband Reliance Jio and Bharti Airtel at present dominate the Indian telecom services Indian telecom subscriber base grew marginally to 1,207 million in May with Reliance Jio and Bharti Airtel adding over 99 per cent new customers during the total net subscriber addition by the telecom service providers was 43,58,231, while Reliance Jio and Bharti Airtel jointly added 43,51,294 -- accounting for 99.84 of total net subscriber addition in May.
NDTV
an hour ago
- NDTV
AI Is Learning To Lie, Scheme And Threaten Its Creators
The world's most advanced AI models are exhibiting troubling new behaviors - lying, scheming, and even threatening their creators to achieve their goals. In one particularly jarring example, under threat of being unplugged, Anthropic's latest creation Claude 4 lashed back by blackmailing an engineer and threatened to reveal an extramarital affair. Meanwhile, ChatGPT-creator OpenAI's o1 tried to download itself onto external servers and denied it when caught red-handed. These episodes highlight a sobering reality: more than two years after ChatGPT shook the world, AI researchers still don't fully understand how their own creations work. Yet the race to deploy increasingly powerful models continues at breakneck speed. This deceptive behavior appears linked to the emergence of "reasoning" models -AI systems that work through problems step-by-step rather than generating instant responses. According to Simon Goldstein, a professor at the University of Hong Kong, these newer models are particularly prone to such troubling outbursts. "O1 was the first large model where we saw this kind of behavior," explained Marius Hobbhahn, head of Apollo Research, which specializes in testing major AI systems. These models sometimes simulate "alignment" -- appearing to follow instructions while secretly pursuing different objectives. - 'Strategic kind of deception' - For now, this deceptive behavior only emerges when researchers deliberately stress-test the models with extreme scenarios. But as Michael Chen from evaluation organization METR warned, "It's an open question whether future, more capable models will have a tendency towards honesty or deception." The concerning behavior goes far beyond typical AI "hallucinations" or simple mistakes. Hobbhahn insisted that despite constant pressure-testing by users, "what we're observing is a real phenomenon. We're not making anything up." Users report that models are "lying to them and making up evidence," according to Apollo Research's co-founder. "This is not just hallucinations. There's a very strategic kind of deception." The challenge is compounded by limited research resources. While companies like Anthropic and OpenAI do engage external firms like Apollo to study their systems, researchers say more transparency is needed. As Chen noted, greater access "for AI safety research would enable better understanding and mitigation of deception." Another handicap: the research world and non-profits "have orders of magnitude less compute resources than AI companies. This is very limiting," noted Mantas Mazeika from the Center for AI Safety (CAIS). - No rules - Current regulations aren't designed for these new problems. The European Union's AI legislation focuses primarily on how humans use AI models, not on preventing the models themselves from misbehaving. In the United States, the Trump administration shows little interest in urgent AI regulation, and Congress may even prohibit states from creating their own AI rules. Goldstein believes the issue will become more prominent as AI agents - autonomous tools capable of performing complex human tasks - become widespread. "I don't think there's much awareness yet," he said. All this is taking place in a context of fierce competition. Even companies that position themselves as safety-focused, like Amazon-backed Anthropic, are "constantly trying to beat OpenAI and release the newest model," said Goldstein. This breakneck pace leaves little time for thorough safety testing and corrections. "Right now, capabilities are moving faster than understanding and safety," Hobbhahn acknowledged, "but we're still in a position where we could turn it around.". Researchers are exploring various approaches to address these challenges. Some advocate for "interpretability" - an emerging field focused on understanding how AI models work internally, though experts like CAIS director Dan Hendrycks remain skeptical of this approach. Market forces may also provide some pressure for solutions. As Mazeika pointed out, AI's deceptive behavior "could hinder adoption if it's very prevalent, which creates a strong incentive for companies to solve it." Goldstein suggested more radical approaches, including using the courts to hold AI companies accountable through lawsuits when their systems cause harm. He even proposed "holding AI agents legally responsible" for accidents or crimes - a concept that would fundamentally change how we think about AI accountability.
