When you hear the phrases "artificial intelligence", "machine learning" or "autonomous systems", what images do they conjure up?
You might be imagining a world with endless possibility. The eternal optimist?
Or perhaps a dystopian future - a robotic society. The not so optimistic?
Or maybe something in-between? The balanced view?
Whatever it is, there is little doubt that artificial intelligence (AI) has entered the mainstream and is not going away just yet.
A car engine - now less familiar than a search engine.
If you use Google to search - and who doesn't - you're already using AI*.
*Search engines typically use machine learning - a branch of artificial intelligence.
The opportunities are vast - but they come, of course, with risk.
What are some of the associated risks that you need to be aware of (for your career and for your organisation's success)?
Let's start with some of the organisational risks:
1. Customer protection: First do no harm. Using AI incorrectly can be catastrophic. Just ask the guys that are developing self-driving cars. AI needs to also be carefully crafted to respect, and protect, privacy.
2. Customer expectations: Matt Turck, a partner at Firstmark (a VC firm with an impressive portfolio) has said that "Customers expect your AI to be superhuman". Customers expect that your products / services just work - consequently, any AI that you use needs to just work.
There is tolerance for human mistakes, but not so much for algorithm errors. We all know how sophisticated the Google search is - but if it doesn't work - we get frustrated.
3. Customer trust: Your customers are growing in sophistication. If you add too much colour to your AI adoption claims - they will likely not translate into better customer outcomes.
Eventually your customers will notice the overstatement and call you on it. The reality is that "calling you on it" may mean talking with their feet - as they exit to competitors that are not exaggerating their claims - humans prefer honesty.
Customer perspective first - in line with our outside-in approach, outlined in this blog.
4. Technology: Traditional technology risk and control practices extend to AI - among them are:
cyber / security protection (e.g. preventing hackers from accessing the AI brain)
change control (e.g. ensuring that algorithm changes are tested)
3rd party oversight (e.g. monitoring API performance, data cache retention and access).
5. Data: Governance in general, but particularly data quality - AI cannot work well with poorly managed data - the lower the quality, the higher the risk of failure.
6. People: Your people are not ready and they either know that they aren't (morale risk) or don't know that they aren't (efficiency risk).
Or your people want to explore, but they are blocked from doing so, and you risk losing them to competitors.
7. Failure to adopt: If you have not yet started to, at minimum, experiment with AI - you are already behind the curve. While there are services and tools that enable acceleration, it takes time to build literacy across the organisation - sustainable adoption does not happen overnight.
And for you - your ability to deliver / success in your career?
Among the many risks - and opportunities - here are some of the key things to consider:
1. Leading self: Have you blocked the possibilities out or are you open to understanding and embracing the inevitable changes?
2. Leading others: Are you enabling your team to understand, explore and adapt?
3. Leading your organisation: How do you ensure that your organisation is preparing? Are you a dissenting voice?
The 10 risk considerations outlined here are clearly not a comprehensive set.
But they could help fill gaps in your AI risk profile, or could perhaps be a starting point.