How Blockchains Can Help Solve AI’s Deepfake Problem

As AI continues to work itself into our every day lives, it is onerous to not see the impression it is already having on practically each sector. Inside the finance trade, for instance, AI is facilitating smarter investments, analyzing market developments and predicting inventory efficiency, finally serving to people and establishments make extra knowledgeable enterprise selections.

Whereas many of the developments with AI are thrilling and proceed to push completely different industries ahead, there are these abusing the know-how for extra nefarious functions. With generative AI, one of many largest dangers that people and organizations want to concentrate on are referred to as “deepfakes.”

You are studying Crypto Lengthy & Brief, our weekly e-newsletter that includes insights, information and evaluation for the skilled investor. Sign up here to get it in your inbox every Wednesday.

Deepfakes are extremely practical digital forgeries produced with AI to control or generate visible and/or audio content material. For instance, a deepfake would possibly contain an AI-generated video exhibiting a celeb participating in actions or making statements that by no means really occurred, comparable to when comic Jordan Peele created a deepfake of Barack Obama to showcase the risk AI-generated know-how might current.

Whereas we might default to believing what we see, such a cast or misleading AI-generated content material is turning into more and more extra widespread. Between 2022 and the primary half of 2023, deepfakes as a proportion of content material within the U.S. elevated nearly 13 occasions from 0.2% to 2.6%, based on a current report from Sumsub Analysis.

Specialists are already involved deepfakes may very well be used to attempt to sway public opinion or affect necessary occasions like elections, with unhealthy actors making an attempt to make use of AI to impersonate elected officers. They’re “utterly terrified” that the upcoming Presidential race will contain a “tsunami of misinformation,” pushed closely by deepfake and deceptive AI-generated content material, one other current report famous. Many view deepfakes’ skill to blur the traces between fact and fiction as a elementary risk to democracies and truthful elections across the globe.

So how can we – as a society – mitigate the prevalence and dangers of deepfakes, in addition to comparable dangers that will emerge as generative AI solely continues to get extra subtle?

Blockchains may very well be the essential know-how we have to assist sort out this difficulty. At their core, public blockchains, comparable to Ethereum, have a number of key options that make them uniquely positioned to ascertain authenticity for content material and knowledge. This contains blockchain’s inherent transparency, decentralized nature and deal with community safety and immutability.

For these unfamiliar, a public blockchain transparently data info in a time-bound method, accessible to all, globally, and with out gatekeeping. This permits anybody to confirm the validity of knowledge, comparable to its creator or a timestamp, making it a supply of fact. Public blockchains are additionally decentralized, eliminating the necessity for a central decision-maker, and lowering the danger of manipulation. This decentralized construction additionally affords excessive community safety by eliminating single factors of failure, and guaranteeing an immutable and tamper-resistant file.

Moreover, blockchains have already demonstrated their skill to authenticate content material. For example, with digital artwork as non-fungible tokens (NFTs), blockchain tech permits anybody to confirm the creator and proprietor of a chunk of artwork, enabling our skill to differentiate between the unique and its potential replicas. This transparency and authentication potential extends to movies, pictures, and textual content, offering necessary foundations for builders to create options and instruments geared at combating deepfakes, comparable to OpenAI’s Worldcoin, Irys and Numbers Protocol.

As AI’s impression on society grows, AI-generated content material and deepfakes will solely turn out to be extra outstanding. Harvard specialists already predict that greater than 90% of content material on-line might be AI-generated sooner or later. To guard in opposition to threats comparable to deepfakes, it is essential we get forward of the problem and implement modern options. Public blockchains, collectively owned and operated by customers, provide promising options like community safety, transparency, and decentralization which may also help in opposition to the problems deepfakes current.

Nonetheless, a lot of the work underway stays in its early levels, and challenges stay with the technical improvement and widespread adoption of blockchain-related protocols. Whereas there isn’t a fast repair, we should stay dedicated to shaping a future that upholds fact, integrity, and transparency, as our society navigates these rising applied sciences (and the dangers they current) collectively.

Be aware: The views expressed on this column are these of the writer and don’t essentially replicate these of CoinDesk, Inc. or its house owners and associates.