Danny Boyle's high-energy drama about Murdoch's early years in UK newspapers has opened at the Venice Film Festival – and its flattering portrayal of the man himself is causing debate. Rupert Murdoch is one of the most influential media figures of the last half-century
IAS officer suspended after arrest over KPSC recruitment exam rowThe Karnataka government has suspended IAS officer Gyanendra Kumar Gangwar after his arrest in a case involving alleged corruption in recruitment conducted by the Karnataka Public Service Commission Published on: Sep 4, 2026
The new film by Irish director Martin McDonagh stars John Malkovich as a 1970s agent wrestling with his conscience and in danger of spilling secrets. It's a spy yarn with real bite. Martin McDonagh loves to make films about two men of different ages swearing at each other in scenic locations
‘I panicked’: Pune founder finds secret WhatsApp group, expects office gossip. Then this happensA Pune founder’s reflection on finding a secret WhatsApp group created by his team members has prompted varied responses on LinkedIn. Updated on: Sep 4, 2026
Where is Michael J. Fox now? Inside actor's life after Parkinson's diagnosis amid video with Harrison FordMichael J. Fox returned to the spotlight with Harrison Ford in “Shrinking” and a Parkinson’s PSA, while continuing advocacy and occasional media appearances. Updated on: Sep 4, 2026
Chandrababu Naidu’s techno-utopian vision for his Indian stateHas Andhra Pradesh’s tech-savvy chief minister bitten off more than he can chew? Published on: Sep 4, 2026, 08:01:36 IST The Economist Prefer HTon Google Share via Copy link As he tells the story, Chandrababu Naidu smiles
‘Half the inmates are drug addicts in Punjab’; HC says requires highest priorityThe bench of acting chief justice Ashwani Kumar Mishra and justice Rajesh Gaur also directed Haryana and Chandigarh to provide similar data about the jail inmates in their respective jurisdictions to check the
Rainfall to continue in HP till September 9: IMDAccording to IMD officials, light to moderate rain is likely at many places over the state with heavy rain likely at isolated places till September 5 Published on: Sep 4, 2026, 08:00:20 IST By HT Correspondent
Demis Hassabis has a plan to harness AI safely
Demis Hassabis has a plan to harness AI safely
HomeNext
Demis Hassabis has a plan to harness AI safely
The Google DeepMind co-founder sets out his vision in an exclusive interview
Updated on: Jul 17, 2026, 21:28:40 IST The Economist Prefer HTon Google Share via Copy link THE WORLD’S artificial-intelligence superpowers are beginning to throw their weight around. The American government’s short-lived decision to withhold foreigners’ access to Fable, a whizzy AI model from Anthropic, reflected not just a desire for control over the technology but also fear of the dangers it could pose in the wrong hands. Now China is said to be considering curbing overseas access to its most advanced models. Yet for AI to be harnessed safely, the industry needs a plan that is coherent and predictable. Sir Demis Hassabis, the co-founder of Google DeepMind, thinks he has one. Sir Demis Hassabis, the co-founder of Google DeepMind, (HT_PRINT) Sir Demis has sketched out proposals for AI regulation in the past. His latest attempt—published online on July 14th and elaborated in an interview with The Economist—is more concrete. The American government, he says, should develop a system for testing the safety of new AI models before they are released. “It’s important that it’s not just an industry body,” he adds. But a regular government agency wouldn’t do either. “It would not be able to move fast enough, or have the right resources.” Instead, Sir Demis suggests taking inspiration from FINRA, the Financial Industry Regulatory Authority, a private agency in America that regulates brokers and stock markets. It is a pragmatic proposal. Gone are the optimistic hopes, still held by some of his peers, of first establishing international consensus: America, he says, must lead the way. Agreement is not required because other countries will fall in line in order to maintain access to American technology and the vast American market. One day that could even extend to China, helping to avert a race to the bottom on safety. Although Sir Demis thinks the scheme should initially be voluntary for model-makers, his expectation is that it will end up mandatory. FINRA, by way of example, has hard enforcement powers delegated to it by the state. Sir Demis’s approach differs from that suggested by Sam Altman, the boss of OpenAI, who wrote an op-ed in the Financial Times earlier this month. Both men agree that the work of the new oversight body would be to establish standards, analyse risks and ensure that only countries that have joined the cause gain access to cutting-edge AI. Mr Altman, however, has called for an international effort, albeit one co-ordinated by America. Both proposals were presented to world leaders at a recent G7 summit in France. Sir Demis’s new proposal is the most detailed to have arrived from an AI bigwig. So far the industry’s senior figures have been more comfortable calling for rules in the abstract than designing regulatory agencies from scratch. Sir Demis, by contrast, has considered many of the practical details. He says the new agency should be funded by the industry, and that much of its immediate attention would be directed towards recruiting a talented team (existing AI-safety bodies, such as those in America and Britain, could offer support in the short term). Sir Demis suggests that the new agency should focus on the labs making the most capable AI models (which includes his own). Earlier proposals from America and the European Union looked to the amount of computing power used to train a model as a rough guide for when oversight was required. Sir Demis instead suggests designating AI models as “frontier” if they meet certain thresholds across a selected set of benchmarks. The creators of those models would then be designated as “frontier labs”, placing extra responsibilities on them. Sir Demis is proud of the “elegant” way that approach sidesteps the question of whether academic or open-source models should be included or not. What matters is their capabilities. Selecting the right benchmarks would be an important function of the new agency. Existing tests of AI models focus largely on commercial capabilities, including software coding and general knowledge. Some organisations have gone further: Britain’s AI Security Institute, for example, has developed a well-regarded test for hacking capabilities which simulates breaking into a power station. But other characteristics, such as how willing a model is to ignore a user’s instructions or pretend to be less capable than it is, are rarely examined in a systematic way that could be used to set a safety threshold. Securing support from within the increasingly fractious AI community will not be easy, though the areas in common with Mr Altman’s proposal suggest it is not impossible. The American government also appears to have been listening, and is said to be close to unveiling its preferred approach. Then will come the hard part. “No proposal would be immune to being executed poorly,” Sir Demis admits. “There are obviously versions of this that could be not very useful.” Yet the Nobel laureate argues that urgent action is required. “Time is short, the hour is late,” he says. “We’ve seen the cyber threat; there’s much more serious threats coming down the line. Something has to be done: this is the time.” To track the trends shaping commerce, industry and technology, sign up to “The Bottom Line”, our weekly subscriber-only newsletter on global business. Home/Technology/Demis Hassabis Has A Plan To Harness AI SafelyTHE WORLD’S artificial-intelligence superpowers are beginning to throw their weight around. The American government’s short-lived decision to withhold foreigners’ access to Fable, a whizzy AI model from Anthropic, reflected not just a desire for control over the technology but also fear of the dangers it could pose in the wrong hands. Now China is said to be considering curbing overseas access to its most advanced models. Yet for AI to be harnessed safely, the industry needs a plan that is coherent and predictable. Sir Demis Hassabis, the co-founder of Google DeepMind, thinks he has one.
Sir Demis has sketched out proposals for AI regulation in the past. His latest attempt—published online on July 14th and elaborated in an interview with The Economist—is more concrete. The American government, he says, should develop a system for testing the safety of new AI models before they are released. “It’s important that it’s not just an industry body,” he adds. But a regular government agency wouldn’t do either. “It would not be able to move fast enough, or have the right resources.” Instead, Sir Demis suggests taking inspiration from FINRA, the Financial Industry Regulatory Authority, a private agency in America that regulates brokers and stock markets.
It is a pragmatic proposal. Gone are the optimistic hopes, still held by some of his peers, of first establishing international consensus: America, he says, must lead the way. Agreement is not required because other countries will fall in line in order to maintain access to American technology and the vast American market. One day that could even extend to China, helping to avert a race to the bottom on safety. Although Sir Demis thinks the scheme should initially be voluntary for model-makers, his expectation is that it will end up mandatory. FINRA, by way of example, has hard enforcement powers delegated to it by the state.
Sir Demis’s approach differs from that suggested by Sam Altman, the boss of OpenAI, who wrote an op-ed in the Financial Times earlier this month. Both men agree that the work of the new oversight body would be to establish standards, analyse risks and ensure that only countries that have joined the cause gain access to cutting-edge AI. Mr Altman, however, has called for an international effort, albeit one co-ordinated by America. Both proposals were presented to world leaders at a recent G7 summit in France.
Sir Demis’s new proposal is the most detailed to have arrived from an AI bigwig. So far the industry’s senior figures have been more comfortable calling for rules in the abstract than designing regulatory agencies from scratch. Sir Demis, by contrast, has considered many of the practical details. He says the new agency should be funded by the industry, and that much of its immediate attention would be directed towards recruiting a talented team (existing AI-safety bodies, such as those in America and Britain, could offer support in the short term).
Sir Demis suggests that the new agency should focus on the labs making the most capable AI models (which includes his own). Earlier proposals from America and the European Union looked to the amount of computing power used to train a model as a rough guide for when oversight was required. Sir Demis instead suggests designating AI models as “frontier” if they meet certain thresholds across a selected set of benchmarks. The creators of those models would then be designated as “frontier labs”, placing extra responsibilities on them. Sir Demis is proud of the “elegant” way that approach sidesteps the question of whether academic or open-source models should be included or not. What matters is their capabilities.
Selecting the right benchmarks would be an important function of the new agency. Existing tests of AI models focus largely on commercial capabilities, including software coding and general knowledge. Some organisations have gone further: Britain’s AI Security Institute, for example, has developed a well-regarded test for hacking capabilities which simulates breaking into a power station. But other characteristics, such as how willing a model is to ignore a user’s instructions or pretend to be less capable than it is, are rarely examined in a systematic way that could be used to set a safety threshold.
Securing support from within the increasingly fractious AI community will not be easy, though the areas in common with Mr Altman’s proposal suggest it is not impossible. The American government also appears to have been listening, and is said to be close to unveiling its preferred approach. Then will come the hard part. “No proposal would be immune to being executed poorly,” Sir Demis admits. “There are obviously versions of this that could be not very useful.”
Yet the Nobel laureate argues that urgent action is required. “Time is short, the hour is late,” he says. “We’ve seen the cyber threat; there’s much more serious threats coming down the line. Something has to be done: this is the time.”
To track the trends shaping commerce, industry and technology, sign up to “The Bottom Line”, our weekly subscriber-only newsletter on global business.
Source: HindustanTimes
Related Posts: Sir Demis Hassabis How to harness the full power of Claude FIFA Boss Gianni Infantino Looks To Harness Fighter jet crews parachute safely after collision at US air show Google brings Gemini to the cloud safely with its new Private AI Compute Artemis crew home safely after completing historic mission to the Moon How will Gaganyaan astronauts return safely to earth Crew evacuated safely from burning ship in San Pedro How can parents safely support children’s phone usage Two more India-flagged LPG tankers safely transit Strait of Hormuz
Bethlehem Kudumba Unit writer Kiran Josey reacts to Ohm Shanti Oshana comparison, working with Nivin Pauly | InterviewBethlehem Kudumba Unit, the Malayalam film co-written by Kiran Josey and Girish AD, opened to glowing reviews and has emerged as a huge box office success. Published on: Sep 4, 2026
10 hours ago
The sheer size and noise of data centres are what people tend to notice first. Some say they sound like a jet plane about to take off, others complain of a high-pitched whine. When Lucy* puts her child to bed, she hears a constant humming. "I'll hear it and think, 'What is that? What is going on?'
10 hours ago
The Supreme Court stated the Bar Council of India lacks disciplinary power over law students. Such authority rests solely with the educational institutions where students are enrolled. The Bar Council can only act after a student graduates and registers as an advocate
10 hours ago
Why get an MBA when you can watch a sports documentary?Business schools are running out of management ideas, but dressing rooms are full of them Published on: Sep 3, 2026, 21:27:47 IST The Economist Prefer HTon Google Share via Copy link The world’s top business schools have long debased
10 hours ago
India attracted $136.3 billion in foreign exchange inflows through special measures, the Reserve Bank of India said on Wednesday. The inflows boosted the country’s foreign exchange reserves and strengthened the central bank’s efforts to defend the rupee against external shocks amid global
10 hours ago
Chelsea have had a few remarkable transfer summers since their forced sale from Roman Abramovich's control in May 2022. Once again they were big spenders, outlaying £349m on 11 signings such as Morgan Rogers, Maxence Lacroix and Emiliano Martinez
10 hours ago
ISRO has begun the countdown for a significant satellite launch. This mission will deploy India's first imaging satellite from geosynchronous orbit. The GSLV-F17 rocket is scheduled for liftoff early on September 4. This marks the 19th flight for India's Geosynchronous Satellite Launch Vehicle
10 hours ago
Several users across the globe have reported issues accessing and using major generative AI platforms, according to the outage-tracking website Downdetector. The website includes OpenAI’s ChatGPT, Anthropic’s Claude and XAI’s Grok According to data from Downdetector
10 hours ago
Salman Khan’s Bigg Boss Season 20 is set to launch on September 6, 2026, streaming at 9 pm on JioHotstar, followed by a repeat telecast on Colors TV at 10:30pm. Salman Khan is back, and so is the annual guessing game over who’s walking into the Bigg Boss house — and how much the superstar is
10 hours ago
September is shaping up to be a crucial month for US immigration applicants, with changes affecting green cards, H-1B workers, international students and immigrant visa applicants taking effect or moving closer to implementation. Several developments have firm dates, while others remain proposals
10 hours ago
The India findings of the Microsoft Work Trend Index (WTI) 2026 paint a rosy picture about AI adoption in the country. However, it puts the onus on leaders to scale up the adoption for it to bear fruit at a larger scale. India is emerging as the biggest hub for “Frontier Professionals”
10 hours ago
When Bangladesh’s Foreign Minister Khalilur Rahman told parliament last week that Dhaka would view an invitation to join the Mecca Joint Defence Agreement positively, he framed the country's possible participation as “a measure for the self-defense of the global Muslim community”
10 hours ago
Canva’s 100 feature milestone stitches Docs and more deeper into the workplaceCanva’s latest updates underlines a bigger ambition, to the fragmented workplace tools, bringing everything into a connected Visual Suite proposition Published on: Sep 3, 2026
10 hours ago
Op Sindoor planner tied to choosing terror targets in Pakistan: Who is Ram Temple CEO Jeetendra Mishra?Jeetendra Mishra accompanied Prime Minister Narendra Modi to the Adampur air base in Punjab after Operation Sindoor. Updated on: Sep 3, 2026, 08:00:02 IST By Rahul Singh
1 days ago
Banks’ 'Certificate of Deposits' rates dip as FCNR flows hit $127 billionThe RBI’s scheme for attracting foreign currency deposits from NRIs has boosted banking-system liquidity, reducing banks’ need to raise funds through CD's. Updated on: Sep 3, 2026
1 days ago
WASHINGTON — The Trump administration is wading into a legal fight between OpenAI and the New York Times, backing the ChatGPT-maker's argument in a closely watched case over how it builds its AI technology by ingesting and processing millions of news articles
1 days ago
John Ternus has officially taken over as Apple CEO, succeeding Tim Cook, who has transitioned to Executive Chairman. In his first letter to the employees, Ternus thanked everyone for their support and expressed his appreciation for Cook's leadership
1 days ago
IAF veteran, Delhi businessman, lawyer, RSS worker: Meet Ram Temple Trust’s new facesJeetendra Mishra, the Ram Temple trust’s first CEO, is an IAF veteran. He was also among the senior planners of Operation Sindoor. Updated on: Sep 3, 2026, 07:42:26 IST By Pawan Dixit | Edited by Priyanshu
1 days ago
Controversy is swirling around Rachel Cusk's new novel and its movie-star protagonist named "M". Why the fuss – and could the character be based on Natalie Portman? From Truman Capote mining the secrets of his high society friends for his writing to The Devil Wears Prada's Miranda Priestly
1 days ago
Apple is expected to unveil the iPhone 18 Pro series on September 9, with pre-orders reportedly opening later than usual on September 12 and sales beginning on September 18. Apple is expected to unveil its next-generation iPhone 18 Pro range at a special event on September 9
1 days ago
Indian envoy to US Vinay Kwatra rings Nasdaq bell, hails economic ties between nationsFirst opened in 1971, the Nasdaq is one of the world’s two largest stock exchanges by market capitalization alongside the New York Stock Exchange. Updated on: Sep 2, 2026
1 days ago
Lego Skylines, John Ternus, and assessing Qubo R700 before the pollution spikeBeyond the regular noise, a closer look at personal technology. 03 September 2026. Published on: Sep 3, 2026, 08:00:22 IST By Vishal Mathur Prefer HTon Google Share via Copy link Opening thoughts. There was a time
1 days ago
India retains status as fastest-growing large economy with latest GDP growth of 7.8%: KwatraIndia retains status as fastest-growing large economy with latest GDP growth of 7.8%: Kwatra Published on: Sep 3, 2026, 07:22:47 IST PTI Prefer HTon Google Share via Copy link New York, India's 7
1 days ago
Did Travis Kelce’s ex Kayla Nicole really ‘expose’ him? Viral cheating and $5M claims debunkedThe viral Kayla Nicole and Travis Kelce claims raise questions about cheating, $5 million, and what Nicole actually said. Published on: Sep 3, 2026
1 days ago
It is a one-time voluntary disclosure scheme contained in Chapter IV (sections 130 to 144) of the Finance Act, 2026. It enables eligible taxpayers to declare certain undisclosed foreign assets, undisclosed foreign income, or undeclared foreign assets, on payment of a specified tax or fee
1 days ago
For almost two decades, Lady Victoria Beckham has been trying to prove that she is more than simply Posh Spice. Now, her fashion and beauty business has reached a milestone many doubted it would achieve: profitability. Victoria Beckham Holdings announced an operating profit of £7
1 days ago
Spotify will not ban AI-generated music but will focus on preventing impersonation. The company introduced AI Credits for artists to disclose AI usage in songs. New AI Persona badges will identify profiles that appear to be AI-generated. This approach aims to inform listeners and protect genuine
1 days ago
