Neural Smash founder Dan Bradley looks at the collective cost of not knowing what’s real and what’s not thanks to AI.
In the last year or so a niggling little voice has crept into the collective consciousness. “Is it real?” it asks as you scroll. It’s a half-second of doubt that’s instinctively applied to almost everything and it never used to be. We’re each running a background
check on every image and video that presents itself. Cognitive scientists have a name for this: verification burden, the mental work of judging something you didn’t create, and the reason it’s so tiring is that the fakes now look completely plausible.
You can’t just skim and trust, you have to stop and think. Multiply that by a thousand posts a day and you have an entire profession’s worth of unpaid labour landing on every person with a phone. We live in a post-truth age, where disinformation has become a daily feast laid on by the likes of Donald Trump. AI is one more float in this carnival of chaos. More worrying than the scams, it’s not just the fake we doubt – we’ve started doubting the real. A genuine photo prompts the same flicker of suspicion
as a generated one. A real video of a real moment gets dismissed as probably AI. And if the doubt keeps growing with nothing answering it, people don’t settle on careful scrutiny, they arrive at trusting nothing at all.
That should concern the funeral industry more than most because authenticity is everything. Think about what passes through a family’s social media feed in the days around a death. The family has to verify tributes as we enter a world where generated
memorials (be it bringing photos to life or reimagining memories) and more recently AI scams, catch us at our lowest and most vulnerable. Your business is built on trust so it would be prudent to stay away from using or producing synthetic images to protect your customer confidence. You can’t fix the doubt by asking people to look harder because we’re already looking as hard as we can.
The answer is to take the burden from the individual and return it to the infrastructure. Encouragingly, behind the scenes that’s finally starting to happen. From August, the European Union’s AI Act will require AI-generated content to carry machine-readable marking, so it can be detected as synthetic. Handset makers are moving too. Every photo taken on Google’s Pixel 10 now carries a tamper-evident record of how it was made, and Google flags known AI images in search with a mark you can tap for the full story. It’s not perfect and the technology is still emerging, but think of it as a nutrition label for what we see. Consuming the wrong
information can be just as detrimental to our brains as consuming the wrong food is to our bodies.
Right now, verifying if something is real or not rests on the tired, the busy and the grieving. It should belong with the platforms serving the content and the tools that made it. I’m optimistic that before long we will have better, tamper-proof cryptographic
proofs of content creation. The sooner that happens, the sooner people can go back to extending the small, generous thing we used to give freely and barely noticed when we did: the benefit of the doubt.