“The world is not yet ready”
With this statement, Alberto, the pseudonymous creator of an artificial intelligence image editor, withdrew his app from sale. The app had gone viral after it received media coverage from Vice’s Motherboard technology blog that was subsequently picked up by other outlets — first online and then TV, radio and print media.
Recovering from the financial damage of a recent failed start-up, Alberto had finally struck product-market fit. It was the type of success founders dream of, with the explosion of traffic and sales taking down his website, he quite literally could not keep up with demand.
Alberto had created something customers wanted, he had identified his capabilities to build his business and despite the fact, as Alberto acknowledged, users could achieve the same results in competing programs like Adobe photoshop, Alberto was cornering the market in his niche.
According to Kenichi Ohmae’s 3 C’s model, Alberto had his strategic triangle nailed.
Why, after so much failure and on the brink of success, did Alberto pull the plug?
Alberto’s app was built using generative adversarial networks (GANs), a class of machine learning described by Facebook’s Artificial Intelligence Chief Yann LeCun as “the coolest idea in machine learning in the last twenty years”. GANs have been used for tasks as diverse as improving the image quality of remastered video games and aiding astronomical research into dark matter. They are also able to generate photo-realistic images of artificial humans.
It is this last capability that Alberto’s app relied on. Unlike his noted competitor, Adobe Photoshop, Alberto had created a single purpose image editor. Building on the emergence of Deep fakes as a means of creating photo-realistic fake images and videos of politicians and public figures, Alberto’s app, DeepNudes, gave users the ability to create fake nude images of women in a click.
The app’s algorithm had been trained using a set of thousands of adult images and worked exclusively on transforming photographs of women. The coverage by Vice brought not only customers, but also complaints about the damage and abuse Alberto’s app could unleash on the world.
Within the day, Alberto had withdrawn his app from sale.
How did we get here?
Alberto’s story is not a new one. A creator, inventor, founder or executive hits upon an idea. It’s a product that has unmatched product-market fit, customer orders outstrip supply and suddenly, out of seemingly nowhere, that rocket ship growth explodes in moral controversy.
Sometimes that ethical flaw is enough to bring the company down, like in Alberto’s case, sometimes they stumble and recover, like in the case of Facebook and Cambridge Analytica. Shy of monopolies, those that survive are rarely the same and that explosive competitive advantage is no longer as sustainable as once thought.
These are the big stories and these big explosions get the bulk of attention. Yet on a smaller scale these stories are playing out every day in ordinary companies.
The Case for Conscience
Kenichi Ohmae’s 3 C’s model has been a staple framework of strategy practice for nearly 40 years. Ohmae posits that aligning customer needs, company capability & competitive positioning creates the strategic triangle of sustained competitive advantage.
Yet as we have seen, whether you are a major bank, tech giant or solo app developer, focusing solely on aligning your customer, company and competitive strategy leaves you vulnerable. While a triangle may make for the strongest of shapes, it does not make for the strongest of strategies.
By adding conscience to your strategy framework, you inoculate your firm from the big missteps that drown and quagmire the firms that do not. Conscience helps firms refrain from falling into the trap of short-term tactical advantage that undermines sustained competitive advantage. When faced with the endless potential of artificial intelligence, ethical strategy serves to help us discern what would should do from the vast ocean of what we could do when it comes to leveraging this technology.
Part of the reason people both readily agree to the importance of ethics in strategy and just as readily avoid incorporating ethics into strategy comes from the diversity and ambiguity of ethical standards of people and groups. Every person wants organisations to behave in line with high ethical standards, specifically their own.
Adding conscience to your strategy framework can often cause people to get lost in the wilderness of philosophical debate and moralising. This is generally not conducive to your overall purpose of defining your strategy.
Instead of seeking to define your organisation’s ethics, it is more effective to view your strategy through the lens of diverse ethical standpoints. By taking a white hat approach and stress testing your strategic alignment to conscience you can identify how you might adapt your strategy or pre-empt failure if you have built your organisation around a fatal flaw.
Debating whether your organisation is ethical is a futile exercise that will draw either universal agreement or cause deep rifts. When developing your strategy, it is more useful to establish that there are a diversity of ethical view points and that you will almost certainly run afoul of some of them.
5 Questions for Ethical AI Strategy
From this base, seek to consider these 5 questions:
- Under what ethical perspectives might our organisation fall short?
- How might we fall short of these perspectives?
- What risks and costs do we face by falling short of these perspectives?
- What competing ethical perspectives does our strategy support?
- How might we improve our alignment between Customer, Company, Competition and Conscience to create and maintain a sustained competitive advantage?
By considering each of these questions you can begin to strength the ethical alignment of your organisation by identifying your weaknesses and strengths, minimising liabilities and building on opportunities. Rather than debating the ethics of your strategy, you instead consider the strategy of your ethics in the context of risks and opportunities afforded.
Conscience in competition
While ethical considerations can be viewed by some as a burden, just as eschewing ethical alignment creates a liability for your firm, understanding your ethical positioning can be a powerful asset. This means not only understanding under which perspectives you are sound and unsound, but also those under which your competitors are so.
A contemporary example of this is the way Tim Cook has waged the privacy crusade to create competitive space between Apple and the rest of big tech. At a time when Apple has been falling behind competitors like Google, Amazon and Microsoft in cloud computing and voice assistants, Cook has made user privacy central to Apple’s strategy.
You aren’t choosing between a “smart” smart assistant and a “dumb” smart assistant, you are choosing between your ‘fundamental human right to privacy’ and the sale of your personal data. With both sides of US politics expressing distrust in big tech firms, this has been an effective alignment of customer, company, competition and conscience for Apple.
Such has been the impact of Cook’s privacy crusade that in their developer conferences this year, Facebook, Google and Microsoft all sought to par Apple’s advantage and declare their own commitments to privacy. Yet with businesses largely built on knowing everything about you, Facebook and Google have been limited in their ability to defend their position.
By contrast, this failure of alignment between company and conscience is why Google’s response has fallen short. The retort of Google Chief Sundar Pichai in the New York Times that ‘Privacy Should Not Be a Luxury Good’ muddles through a contemplation amounting to ‘But what even IS privacy…?’ and jibes at Apple for being expensive. Pichai effectively made Cook’s point, if you want privacy take your business to Apple.
The Sense Check
One final step you can take to promote the ethical alignment of your strategy is to sense check your personal ethics. For Alberto there were clear warning signs that what was a booming business with clear product-market fit would soon be doomed. Needing a pseudonym to promote your business is usually not a great indication that you are comfortable.
When you are formulating or receiving your A.I. strategy, ask yourself:
‘Is my conscience clear on this strategy?’
If the answer to that question is not yes, then you will have to change your organisation. If you can’t change your organisation, change your organisation.
If you liked this article, join the newsletter: https://tinyletter.com/OnStrategy