Deepfake content is a fascinating topic to discuss. On the one hand, it's an example of where technology progress is right now, and what it could be in a short time. On the other hand, are current use cases for deepfake technology the best we can do? It may be fascinating to see Harrison Ford recreated seamlessly as a 40-year-old, but it's not nice to see the technology being used for underage sexualised content.
As ever, read my scribbled preparation notes for the recording of episode 31 of Big Tech Little Tech. Rick and I discuss a few things, and far more than the audio version. If you're a Patreon subscriber, you can spend £2 a month to get extended video versions and transcripts.
What is pig butchering?
These are crypto scams.) It costs a lot to scam people – sometimes money, but oftentimes it’s a labour-intensive process to fatten somebody up for a crypto scam. Imagine removing those labour costs with deepfakes that draw you in and develop a relationship with you, before destroying your life.
When deepfakes become harder to notice, what does this mean for trust?
When text-to-voice becomes as good as voice-to-voice – when it’s able to simulate inflections, comedic timings, characteristic phrasing and so on – is when things may change dramatically.
They could establish clear policies regarding communication and verification procedures.
They could provide guardrails to protect our identities and the identities of others.
What are the challenges for deepfake developers?
Engendering trust.
Convincing us that we need convincing simulations that will enrich our lives outside of popular entertainment such as Mission Impossible, Terminator and Star Trek.
Accents change as we move around.
Language simulation. Star Trek was ahead of the times! Imagine going to a foreign country and being able to converse in Spanish in real-time.
People want to skip reading.
People want to skip writing.
People will want to skip learning if they can fake it enough to get by.
Can we be optimistic? Can blockchain help in proving provenance?
Terminator phone calls!
Voicemail messages.
iOS 17 Personal Voice.
Being initially sceptical about something may be a good thing in the long run. For instance, if we see something, and take the time to think about what we’ve seen, instead of have an immediate reaction, this may help us slow down and consider what’s real and what’s not real. In doing so, we may not have so many hot takes on things as much as we did in the past.
Sam Harris argues that we’re not good at being naturally skeptical. It may take software that runs in the background to determine if the information we’re consuming is authentic or true.
Some people may not want to opt in to being skeptical because it’s more seductive to have your biases confirmed.
In future, more people will know and understand what the meaning of Turing Test is.
Thanks again for reading. Perhaps my notes might inform something you're putting together. If so, I'd love to know more. Send me a link to your work, so in turn we can learn from one another. Until then, stay safe, keep smiling and enjoy the details.
