Days after half a billion Asians went to the polls, Big Tech promises to counter 2024 election misinformation
- Reference: 1708320494
- News link: https://www.theregister.co.uk/2024/02/19/big_tech_election_misinformation_accord/
- Source link:
Signatories to "The Tech Accord to Combat Deceptive Use of AI in 2024 Elections" include Adobe, Amazon, Arm, Google, IBM, Inflection AI, LinkedIn, McAfee, Meta, Microsoft, OpenAI, Stability AI, TikTok, Trend Micro, and X.
Accord members unveiled their effort on February 16 by isusing a [1]statement [PDF] about how the members developed "a set of commitments to deploy technology countering harmful AI-generated content meant to deceive voters."
[2]
The group hopes to combat "AI-generated audio, video, and images that deceptively fake or alter the appearance, voice, or actions of political candidates, election officials, and other key stakeholders in a democratic election." False information about when, where and how to vote is also in the group's sights.
[3]
[4]
The crew acknowledged that this year a record number of people will go to the polls – more than four billion people in over 40 countries.
They don't mention that half a billion have already gone to the polls this year. Indonesia and Pakistan, with populations of over 270 million and 230 million respectively, voted on February 8 and 14. Indonesia's government warned of deepfakes influencing its election, while in Pakistan former prime minister Imran Khan [5]campaigned using deepfakes – he was jailed before the election and could not appear in the flesh.
[6]
Among the other 3.5 billion eligible to vote this year are many of India's 1.4 billion citizens, the US's 330-million-plus, and the UK's 65 million. Those three nations are far bigger customers for big tech's wares than Indonesia or Pakistan.
[7]AI political disinformation is a huge problem – but harder to fight than ever
[8]'I'm sorry for everything...' Facebook's Zuck apologizes to families at Senate hearing
[9]Pakistani politician deepfakes himself to deliver a speech from behind bars
[10]Meta: If you're in our house running AI-massaged political ads, you need to 'fess up
Accord signatories adopted a core set of related principles – like tracking the origin of election-related content and supporting public awareness campaigns. They also pledged to "work collaboratively on tools to detect and address online distribution of such AI content, drive educational campaigns, and provide transparency, among other concrete steps."
Other commitments include a promise to assess their products to understand what kind of risks they pose and detect distribution of content on their platforms.
Specific suggestions include developing classifiers like watermark or signed metadata to certify the authenticity of content, allowing users to mark content as AI generated or ingesting open standards based identifiers created by AI production companies in order to detect it, and providing transparency to the public on policies.
While those seem like concrete actions, the eight listed commitments overall are vague – like "seeking to appropriately address this content detected on their platforms," and "fostering cross-industry resilience."
[11]
The accord also appears to bill AI as a possible savior to the problem it created. It [12]notes [PDF] that "AI also offers important opportunities for defenders looking to counter bad actors," and offers rapid detection of deceptive campaigns as well as the ability to quickly take on multiple languages as ways it can help scale a defense.
However, the group did not commit to using AI in these capacities.
The authors of the Accord also offer no timeframe for delivery on any of the plan. The pledge goes no further than current policies for many of its members – and in many cases promises less.
OpenAI has already [13]declared that building GPTs for political campaigning and lobbying was prohibited, as is using its models to power any chatbots that impersonate real people, businesses or governments, or meddle in any democratic processes.
Last year Meta [14]decided to require advertisers to disclose AI-generated or digitally altered content – which frankly hasn't kept misinformation off the social network.
So it could all be a lot of noise that achieves very little. Just like almost every other attempt to clean up social media. ®
Get our [15]Tech Resources
[1] https://www.aielectionsaccord.com/uploads/2024/02/Press-Release-AI-Elections-Accord-16-Feb-2024.pdf
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2ZdOJO6kj@KBlRikOhxJqBAAAAQw&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZdOJO6kj@KBlRikOhxJqBAAAAQw&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZdOJO6kj@KBlRikOhxJqBAAAAQw&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[5] https://www.theregister.com/2023/12/19/ousted_and_isolated_pakistan_pm/
[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZdOJO6kj@KBlRikOhxJqBAAAAQw&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[7] https://www.theregister.com/2024/01/17/ai_political_disinformation/
[8] https://www.theregister.com/2024/01/31/senate_social_media_zuckerberg/
[9] https://www.theregister.com/2023/12/19/ousted_and_isolated_pakistan_pm/
[10] https://www.theregister.com/2023/11/29/meta_ai_ad_disclosure_requirement/
[11] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZdOJO6kj@KBlRikOhxJqBAAAAQw&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[12] https://www.aielectionsaccord.com/uploads/2024/02/A-Tech-Accord-to-Combat-Deceptive-Use-of-AI-in-2024-Elections.FINAL_.pdf
[13] https://www.theregister.com/2024/01/17/ai_political_disinformation/
[14] https://www.theregister.com/2023/11/29/meta_ai_ad_disclosure_requirement/
[15] https://whitepapers.theregister.com/
Big Tech at it again
This is about as convincing as an NRA meeting concluding that school shootings have to stop.
Re: Big Tech at it again
Yup. It's all just mouth noises and beautiful PR until something actually technical is put on the table.
Then it becomes real and we find out if the PR was worth it (it probably won't be).
Follow the Money
It's the advertisers that need to be signed up. They'll spend their money where they get the clicks and if those clicks move to other sites where AI generates popular content, whether true or not, then that's where the money will go. When the current big players start losing ad revenue to nutter-AI sites then they'll forget all about this accord and do whatever they can to get it back.
McAfee
Who the hell are they?
Trellix ...
Thankfully Google have already unleashed their AI on the problem....
Apparently we can no long have a trump suit in Bridge (the card game) in store content, because well... AI is involved so....
Far too late
Most of the usual suspects in the 2024 US POTUS election are already crying 'it is rigged' and there has only been a few primaries. One is saying that because he believes (wrongly) that he won that state in 2016 and 2020 and because he didn't get 95% of the vote, it was rigged against him. That margin is the sort of margin that Putin will win by in his election this year.
As for AI, those same usual suspect are already crying out that their opponents are using Fake images generated by AI in the ads against him and his cult.
This is as they say, only the tip of the iceberg and the races are just warming up at this stage. IT will get nasty and no matter what this cabal of companies might publicly say, when it come to clicks, hits and ad revenue, all these platitudes will go out the window as fast as DJT's claims that he has never ever done anything wrong. I wonder when we'll see some of his US opposition start taking a falls out of windows? His boss seems to excel in that.
Yes I look forward to this latest greatest initiative by the lying bastards in those corporations. It should be just as successful as all the other great things they have done for the web in the past. How about this one, every item produced by AI is marked as such. In bold unmistakable letters and in the case of video watermarks that cannot be removed. Like should have been done right from the start. But that would have been too simply actually taking some responsibility to ensure your new hype train to scam the investors for billions more was not going to cause problems, it might never have left the station then. Oh and I thought we were supposed to be using less resources to save the planet, just where was this factored into the decisions to massively increase them by using this technology for pretty much zero gain for society in its use.
I have a feeling one or two of those signatures had a little caveat at the bottom saying "unless it helps Trump or Putin, in which case it's totally fine with us". Naming no names, of course...
Just ban Australians
From owning all your media companies
Humans have tried to stop humans communicating with humans for tens of thousands of years. It has never worked.