Doping the Peloton - does fake matter if everyone thinks it’s real?

“Looking great there, Arnold1970! Now let's keep on pushing and smash you personal best!” Once the reserve of a personal training session, now these kind of motivations are available from the comfort of your own home.

With a whole range of new fitness machines on the market, bringing the full experience of a tutored gym class to your front room, millions of people are subscribing to gain access to live, interactive classes where you are kept engaged in a way that only that personalised experience can deliver.

It's a great sell, from those with hectic lives, others living too far from a gym to be convenient and even parents looking to fit in a quick workout while children sleep. The "home gym" is hardly new, but what has been lacking until now is that vital aspect of motivation. Gaming the home gym experience is the missing jigsaw piece that previously resulted in expensive exercise equipment moved from the living room to the car boot sale.

However, despite the fairly steep cost of entry [Peloton bikes for example retailing just shy of £2,000] coupled with the monthly subscription, the company continues to make a loss with profitability not currently forecast until 2023.

It's understandable though, considering the company only launched in 2013 via a Kickstarter campaign. There's the R&D that went into the bike design and construction, followed by the need for large-scale manufacturing, stores with commercial and staffing overheads, music rights for the class soundtracks and of course all the IT systems and software development needed to ensure you have that "live" experience of the class.

And then there's the instructors.

Instructor salaries are an overhead too, with these fitness trainers nearing "celebrity" status [in the right circles] there is a demand for talented individuals to lead these classes and with the inevitable rise of competitors in the market, there is of course the issue of retaining the best talent.

But what if that was a problem that could be made to go away, along with the current financial overhead?

In the age of "deepfakes" where anyone with an internet connection and a webcam can make the President of the United States start insulting entire nations.. Or maybe a better example of replacing every person in a TV show with Nicholas Cage. This technology is already at a point where videos that appear online are taken seriously for a short time before being debunked..

Scary stuff, as are the potential opportunities for services such as the Peloton "live" instructors to apply Machine Learning and Artificial Intelligence, combined with a scripted workout session in order to create "Real" Robin.

A clone of their most popular instructor, indiscernible from the real thing, who doesn't get tired, never takes holidays, never calls in sick and is always ready to "crush it".

It's not inconceivable to imagine contracts for those just starting out in the instructor game may start to feature clauses for image rights and "further use" of these images even when the person has left the company. With shareholders ever more impatient for returns on their investments it makes sense that innovation could enter the arena to help keep unpredictable variables [in this case the human instructors] to a minimum.

While all of this would go directly against the reason the phenomena has taken off in the first place [that "personal touch"], if the machines have access to all your previous workouts in the same way a human would, is it not more likely that algorithms and machine learning can detect specific peaks and troughs in your own workout to tailor it to suit your maximum output better than the instructor on screen who is trying to track the 30 new usernames who just joined the session from Idaho?