In the battle to differentiate real from fake…
“It may not seem like much now, but this trial by TikTok is huge news in the fight for widespread disclosure/labelling of GenAI content and deepfakes,” says Henry Ajder, an expert advisor on generative AI, deepfakes, and AI policy.
TikTok appears to be trialling a new way to help creators label their videos that contain AI-generated content. When uploading a video, a new “AI-generated content” labelling option has been introduced for some users, triggering an alert which says the below.
TikTok announced back in March that they would require users to disclose realistic deepfakes and GenAI on the platform to help viewers understand.
But social media is saturated with synthetic and partially AI-altered content. “This is arguably the biggest challenge and is not unique to TikTok,” says Ajder via Linked In.
“AI-powered filters and editing tools have been part of social media for many years. The policy talks about disclosing realistic AI-generated scenes, but what about partially AI-altered or augmented media? Would realistic makeup filters need to be labelled?” he asks
The trial follows a commitment from leading AI companies to build disclosure into their generative tools, as announced by The White House last month.
AI companies including OpenAI, Alphabet (GOOGL.O) and Meta Platforms (META.O) pledged to watermark AI content for safety, President Jo Biden announced.
The companies have made voluntary commitments to the White House to implement measures such as watermarking AI-generated content to help make the technology safer.
In doing so, the firms have agreed to 3 key principles around safety, security and trust. This includes a commitment to prioritising research on societal risks posed by AI systems, internal and external red-teaming of models or systems in areas such as misuse and national security concerns, such as bio, cyber, and other safety areas, as well as developing and deploying frontier AI systems to help address society’s greatest challenges.
FURTHER READING:
- OpenAI, Google, others pledge to watermark AI content for safety, White House says, via Reuters>
- White House secures voluntary pledges from Microsoft, Google to ensure AI tools are secure, via CNBC>
Technology lawyer, Raymond Sun, says that while the commitments don’t address all grey area issues in AI (e.g. intellectual property issues relating to training datasets), they are still a good starting point to set a base level of consumer trust and market expectations around AI safety (especially while the US is still working on a federal AI regulatory framework).
He explains: “While I’ve seen some cynical views/commentary from others on whether these commitments are genuine, the fact that this was a highly publicised announcement (even President Biden gave a televised speech on it) probably creates some accountability as the participating companies are now under public scrutiny to fulfil their commitments.
“In any case, we’re seeing more innovative ways to use AI regulation. Unique/creative approaches have ranged from Russia’s regulatory sandboxes to Singapore’s AI governance toolkits to Japan’s regulatory notices to US tech company commitments. Hard legislation is not the only method,” he adds.
How will AI impact SEO?
“We’ll see more technology changes in the next few years than we’ve seen in the last 50 years”, said President Biden when he made the announcement. And this rings true for its impact on all areas of business.
This new article from Venture Beat raises an interesting point about the impact of AI on traditional Search Engine Optimisation (SEO) methods, a key way for businesses and retailers to get visibility online.
According to founder, investor and longtime industry analyst Jeremiah Owyang, Bill Gates’ vision of personal AI is coming. Gates said personal AI technology could radically alter user behaviours, resulting in people never needing to visit a search website again, use certain productivity tools or shop on Amazon.
And this vision of the future could revolutionise SEO and ecommerce, requiring marketers to overhaul previously successful methods.
Retail expert ShiSh Shridhar says advertisers would be forced to adapt to the rise of AI agents and foundational models, which will change the way users interact with search and e-commerce.
He explains: “Instead of paying for ads on web pages, advertisers will have to pay for their messages to be included in the AI-generated responses that users receive.”
What do you think? Are you already using AI tools to benefit business? We’d love to hear from you.
Visit Matt Haycox Daily for all the latest AI and Technology News>
STAY UP-TO-DATE
For fresh insights, company news and business advice, subscribe to the weekly Matt Haycox newsletter, below.
