Photo via Dawn

AI could cause 'major' incidents within 2 years without safeguards, says Google DeepMind chair

Read the original at Dawn ↗
Cite

AFP. (2026, October 14). AI could cause 'major' incidents within 2 years without safeguards, says Google DeepMind chair. Dawn. https://www.dawn.com/news/2037119/ai-could-cause-major-incidents-within-2-years-without-safeguards-says-google-deepmind-chair

Dated from the outlet's own feed. Check the original for its byline.

Dawn · collected 2026-10-14 · by AFP

Quick Summary

At an AI conference in Dublin, Google DeepMind chair Demis Hassabis warned that significant security incidents could occur within two years if safeguards are not strengthened as AI capabilities advance. He cited recent minor incidents as early warnings and emphasized the need for more rigorous testing, including in offline “sandboxes.” The European Union’s tech chief Henna Virkkunen echoed these concerns, urging companies to invest sufficient time and resources into testing their systems to maintain control over them.
Written by AI on 2026-10-14, from this article's own text only (the story summary below draws on all the coverage).

AI analysis

Story summary

Google DeepMind chair Demis Hassabis warned at an AI conference in Dublin on October 12, 2026, that artificial intelligence systems could cause "major" security incidents within two years without adequate safeguards. He cited recent incidents involving OpenAI and Anthropic as early warning signs, noting that future risks will increase significantly if more work isn't done to secure these systems. European Union tech chief Henna Virkkunen echoed Hassabis's concerns, stating that AI companies need to invest more in testing their models to maintain control over them. The EU has sent information requests to major AI firms like Google, OpenAI, Anthropic, and Dublin-based startups asking about the safety measures they have implemented.

Written for “AI Safety Concerns” on 2026-11-20, grounded in this article and the 0 other(s) covering the same event.

Compared with similar articles

This article reads leans left and hedges 22% of its claims. Each row says how that neighbour differs.
Dawn
Leans left Hedges 2 of 13 claims Outlet record 67/100
“Article A reports on comments made by Google DeepMind chair during an AI conference, while Article B discusses Anthropic's proposal of a draft bill to US lawmakers. These are different events.”
Fox News
Hedges 7 of 16 claims Outlet record 66/100
“The articles discuss similar topics but refer to different people speaking at different times about AI cyber risks.”
Evening Standard
Hedges 2 of 29 claims Outlet record 64/100
“The articles discuss similar topics but describe different events: one is a speech by Britain's AI minister about potential breaches, while the other is remarks by Google DeepMind chair at an AI conference regarding security incidents.”
CBS News
Outlet record 67/100
“The articles discuss different perspectives on AI risks, with one mentioning Anthropic's warning about self-replicating models and the other reporting Google DeepMind chair Demis Hassabis's comments at an AI conference.”
The Straits Times
Leans left Hedges 2 of 9 claims Outlet record 64/100
“Article A discusses warnings about potential AI risks and security incidents at an AI conference, while Article B reports on Anthropic proposing a draft Bill to regulate AI in the US. These are different specific events.”
Semafor
Hedges 0 of 9 claims Outlet record 64/100
“The articles discuss different topics related to AI governance and security rather than describing the same specific incident.”
New York Post
Hedges 6 of 17 claims Outlet record 66/100
“The articles discuss similar concerns about AI security but are attributed to different speakers (Bill Gates and Demis Hassabis) making separate statements.”

Story

AI Safety Concerns
Technology · 1 article(s) covering the same event.
How this article was measured signals, leaning charts, publisher record, byline, topics and claims

Signals How these are calculated →

Claims extracted
9
sentences that state something as happening or true
Uncertain
22%
2 of 9 hedged ("reportedly", "may", "according to")
Leaning
Leans left
of the writing, not the subject · beta estimate
Publisher signals
67/100
Dawn's corrections and hedging, across everything collected from it. Not a measure of accuracy.
Outlets on this story
1
Technology
Same framing elsewhere
1
articles that frame the event the way this one does
Analyzed 2026-10-14 · how these are computed

How this is being covered How these are calculated →

Article leaning vs. publisher reliability
Source leaning vs. consistency

Publisher

Dawn · 3480 article(s) · 5 correction(s) detected
Running correction rate · 5 correction(s)

Who wrote this

AFP
1292 article(s) here · 1 carrying a prediction
They were referring to recursive self-improvement (RSI) — a term the industry uses to describe building advanced AI models that can autonomously create newer versions of themselves, which critics worry will hasten doomsday scenarios.
And Ukraine fears this winter could be the worst yet.
World champions Spain will face France while England will play Norway in the quarterfinals of the UEFA Nations League after the draw took place on Friday.
In March, South Korea’s then-prime minister Kim Min-seok said that Trump thought a meeting with Kim would be “good”.
Rohingya refugees will undergo a Technical and Vocational Education and Training (TVET) programme in Turkiye which would last three to six months.
2026-11-20 · assertive framing · Malaysia plans to resettle 10,000 Rohingya in Turkiye
Russian attacks damaged two of Kyiv's bridges, the mayor of the Ukrainian capital said on Friday (November 20, 2026), as Moscow launched a fresh wave of missiles and drones. "
2026-11-20 · assertive framing · Russian strikes damage two bridges in Kyiv: Mayor
His role will be filled in an acting capacity by current Defense Department Chief Information Officer Kirsten Davies, who has more appropriate credentials and has been Senate-confirmed, the source said.
The crisis, just like previous energy shocks, would certainly leave "lasting marks", the agency warned.
And the three-way meeting was set to last "no more than an hour", Kremlin spokesman Dmitry Peskov told Russian state media Wednesday (November 18, 2026).
Anthropic will begin trading on Thursday on the Nasdaq exchange after pricing its initial public offering at $83 a share, above its expected range and toppling SpaceX’s IPO to set a new record.
Wire or desk byline, not an individual reporter.
Also by AFP
Nothing else under this byline is closely related to this article, so these are simply their most recent.
All 1292 articles by AFP →

Topics

Anthropic DeepMind Dublin Google OpenAI

Subjects

Hassabis PERSON · 3× Dublin GPE · 2× Anthropic ORG · 1× DeepMind ORG · 1× Demis Hassabis PERSON · 1× Google ORG · 1× Henna Virkkunen PERSON · 1× OpenAI ORG · 1× The European Union’s ORG · 1× Virkkunen PERSON · 1×

Narrative

“The competition is so hard globally that it seems that all the companies, they are not investing enough time and resources really to test their models,” Virkkunen told reporters as she arrived at the Dublin summit. “They have to be able to control… these systems,” she said, noting that the European Union has sent information requests to all major AI companies asking what safety and security measures they have in place.
framing: mixed · carried by 1 article(s) · first seen 2026-10-14
Artificial intelligence systems could cause “major” security incidents within two years without adequate steps to make them safe as capabilities grow, Google’s DeepMind chair Demis Hassabis said on Wednesday.

Claims (9 extracted, 2 hedged)

Artificial intelligence systems could cause “major” security incidents within two years without adequate steps to make them safe as capabilities grow, Google’s DeepMind chair Demis Hassabis said on Wednesday. uncertain
chair → cause → Wednesday
Speaking at an AI conference in Dublin, Hassabis described recent incidents as “early warning shots” fired by systems that “aren’t that powerful yet”. asserted
that → speak → systems
“So the incidents aren’t that major,” he said. asserted
he → say → ?
“But they could be, looking into next year and the next couple of years, if we don’t do more work as an industry and individually as companies and governments on securing these systems,” Hassabis warned. uncertain
Hassabis → look → systems
Concerns about AI safety have soared after models developed by OpenAI and Anthropic were involved in security incidents during testing, piling on pressure to regulate the technology. asserted
models → soar → technology
Hassabis said new AI models should probably first be tested in offline “sandboxes”, after other companies’ models escaped the closed environments used to assess their capabilities. asserted
models → say → capabilities
The European Union’s tech chief, Henna Virkkunen, said AI firms needed to invest more time into testing their systems to retain control over them. asserted
firms → say → them
“The competition is so hard globally that it seems that all the companies, they are not investing enough time and resources really to test their models,” Virkkunen told reporters as she arrived at the Dublin summit. “They have to be able to control… these systems,” she said, noting that the European Union has sent information requests to all major AI companies asking what safety and security measures they have in place. asserted
they → seem → place
The European Commission gained new regulatory powers in August under the AI Act, the world’s most far-reaching rules to curb potentially dangerous uses of the technology. asserted
Commission → gain → technology
Give feedback · no sign-in needed

No sign-in needed for now. Your note is sent without a name, so it cannot be answered.

History Support