The Expectations Gap

Last month, the CEO of Grant Thornton, the auditors of Patisserie Valerie, told MPs on the Business, Energy and Industrial Strategy Committee that "We're not looking for fraud", a comment that was taken to be representative of auditors in general. With high-profile failures such as the cake chain and Carillion dominating the media in recent years, this comment sparked some incredulity, thereby highlighting what has come to be referred to as the "expectations gap" in the public understanding of the role of auditors. The purpose of an audit is to independently verify that a company's accounts give a true and accurate picture of its trading status and financial viability, essentially to assure shareholders and debtors (e.g. banks providing financing) that they can rely upon the statutory accounts to accurately gauge risk. This does not extend to assessing the quality of its business - so a firm that is a going concern could still go bankrupt if legitimately anticipated revenue failed to turn up - nor does it include an opinion on a firm's ethical behaviour beyond the conventions of accounting.

The concerns over the audit industry that have emerged since the Enron scandal in 2001 revolve around risk rather than ethics, reflecting a fear among large investors and financial institutions that they are being increasingly exposed by the limitations of the traditional audit. The focus is on firms of the scale of Carillion, not small-fry like Patisserie Valerie, and in particular on businesses with complex intercompany structures that span both the private and public sectors, raising the linked issues of implicit state guarantees and regulatory arbitrage. The expectations gap between what the public expects auditors to be responsible for and what they actually do is not a reflection of popular ignorance but a demand for a more socially-focused role. Though auditors are just business service providers, they are perceived by the public as carrying out a quasi-state function: ensuring that firms are not merely legally-observant in their accounting practices but they are not conducting their business in a manner that might prove damaging to society in the various forms of customers, suppliers and employees.

These two tendencies - the increasing complexity of modern capitalism and the demands for social responsibility - are at root driven by the same trend: the increasing socialisation of capital in the form of firms dependent on institutional shareholdings. The demand for higher ethical standards has long been addressed, if inadequately, by the pabulum of Corporate Social Responsibility (CSR). The demand for greater managerial oversight by institutional investors has been met by legislation on corporate governance, and it is within the latter's ambit that the issue of audit is now being addressed. This means that one problem of the current audit regime, the role of non-executive directors (NEDs) in leading audit committees, is unlikely to be resolved. NEDs have to be clubbable and not ruffle feathers if they are to acquire multiple directorships, with the result that there is little incentive to be hard on either the business or its auditors. Nationalising audit, or at least putting it under the direct control of a state regulator, runs into the problem that it would reduce the power of NEDs and thus call into question the corporate governance regime.

Over and above this there are the traditional objections that a national audit service would represent an excessive incursion by the state into the private sphere, and that the application of a common ethical framework would constrain innovation and enterprise (CSR is optional, after all). More modest calls (e.g. by Labour) for the breakup of the "big four" and the insistence that audit firms should be specialists, rather than loss-leaders used as a platform for upselling, stand a better chance of success, but only if the focus remains on risk management and if state oversight is little more than another "watchdog". What will not be conceded short of nationalisation, and perhaps not even then, is the idea that auditors should assess a firm's social responsibilities. That business should be subject to ethical scrutiny has always been, and remains, anathema. This doesn't mean that ethics will be ignored, but that it will be diverted into the spectacle of personal behaviour, such as "fit and proper persons" tests and disgust at CEO incivility. The Philip Greens and Michael O'Learys of this world will continue to strut the stage, distracting from the theatre itself.

This might all seem a bit dry, but the question of the purpose and organisation of company audits is relevant to the emerging debate over the future management of artificial intelligence, which is nothing is not sexy, at least in its media coverage. This is because both employ the rhetoric of social responsibility but do so in order to preserve established interests. In today's Observer, Will Hutton, who knows a fair bit about capitalist practice but not much about technology, argues for a "smart response" to guard against the sinister abuse of AI. He has a list of measures: "Three seem crucial: maximal transparency and accountability embodied in regulatory oversight and a new Companies Act; methods and digital processes to ensure you own your data and its use; and fast and effective regulation of content. Put another way, we need every organisation deploying digital data to be open and accountable; we need new public interest digital platforms where we can hold our data based on the presumption we own it; and we need another Leveson – a fast and effective mechanism to ensure digital information is not misinformation". This is the classic liberal trifecta: transparency among an elite, secure ownership and civilised censorship.

What is missing in Hutton's analysis is any real understanding of the social context of the technology (this has been a longstanding blind-spot in his oeuvre). For example, he claims that "AI allows the individualisation of your drug treatment and fast and cheap diagnoses of whatever illness you are suffering, along with likely cures". This is nonsense, not least because it is genetic sequencing that might allow such individualisation, not AI, but the wider error here is to fail to see the social reality. Medical research is skewed to the interests of the rich because they have the money to pay for expensive screening and treatments. The idea that AI will produce cheap diagnoses and cures, thereby disrupting a pharmaceuticals industry based on fat profit margins, is naïve. After a gratuitous insult aimed at the Labour Party, which now seems to be obligatory in all Observer comment pieces, Hutton ends with a call to arms that perfectly echoes the fears of an earlier generation of liberals faced with the prospect of universal suffrage: "The task is not to throw up our hands warning that the machines are coming – it is to design a world in which we are their master, not their servant".

The liberal demand that AI be regulated and managed is not motivated primarily by a desire to defend society from its disruptive and destabilising potential, no matter how many times its advocates cite Karl Polanyi, but from the self-interest of existing business sectors. In quoting Shoshana Zuboff's concept of "surveillance capitalism" (a term that is rightly being criticised for having little to do with capitalism), Hutton conflates two completely unrelated technologies. Trawling your Facebook posts to determine what adverts to present you with is not AI, but it is a technique of data personalisation that is eating the lunch of newspapers dependent on advertising. Likewise, claiming that social media is destroying the fabric of civilised life and undermining our politics only makes sense if you cannot imagine either civilisation or politics beyond the frame of elite convention. Both the optimism and pessimism associated with AI reflect expectations of profit more than social benefit. Similarly, the debate over the future of audit is not about preventing fraud or social damage, but mitigating risks to capital.