Update 30 June 2020 the Supreme Court in 2020 SCC 16 has confirmed the earlier finding of invalidity of the first instance court, as reported here.
Update 30 March 2020 Heller is now before the SC, and the British Colombia Supreme Court distinguished it in Williams v Amazon  BCSC 1807, holding as McCarthy Tétrault LLP report that the arbitration clause provided for a refund of costs in many cases. Further, the arbitration could be conducted through flexible options, and as such, this was not a substantially unfair bargain so as to meet the test for unconscionability.
Thank you Christopher Burkett for alerting me to Heller v. Uber Technologies Inc., 2019 ONCA 1. The case is reminiscent of California’s Senate Bill 1241 (review here) and of an article that I co-authored with Jutta Gangsted [‘Protected parties in European and American conflict of laws: a comparative analysis of individual employment contracts]. The starting point of the California, the EU rules, and the Canadian judgment is the same: employees cannot be considered to really consent to either choice of law or choice of court /dispute resolution hence any clause doing same will be subject to mandatory limitations.
Here, an arbitration clause requiring arbitration in the Netherlands of disputes between drivers and Uber was held to be invalid and unenforceable, because it deprives an employee of the benefit of making a complaint to the Ministry of Labour under relevant Ontarian law.
Of note is that the judgment applies assuming the contract is one of employment – which remains to be determined under Ontarian law. Of note is also that the Court of appeal rejected Uber’s position that the validity is an issue for the arbitrator to determine because it is an issue going to the jurisdiction of the arbitrator. Uber invoked the “competence-competence” /kompetenz kompetenz principle (recently illustrated e.g. by the Brisilian Supreme Court in Petrobas) in support of its position.
(Handbook of) European Private international law, 2nd ed. 2016. Chapter 2, Heading 184.108.40.206, Chapter 3, Heading 3.2.5.
The issue under consideration in Citysprint was whether claimant, Ms Dewhurst, a cycle courier, was an employee of Citysprint or rather, as defendant would have it, a self-employed contractor. I am not a labour lawyer but I do have an interest in the ‘gig economy’, peer to peer etc. [Note Google defines (or conjures up a definition of) the gig economy as a labour market characterized by the prevalence of short-term contracts or freelance work as opposed to permanent jobs].
I also have an interest in language and the law. After an employment tribunal in Uber blasted the company’s use of byzantine language, in Citysprint, too, (in particular at 64 ff) the tribunal looks beyond the fog of legalese to qualify the contract for what it really is. A great development.
Update 28 December 2018 the Court of Appeal in  EWCA Civ 2748 has confirmed. Rome I issues no longer featured in the analysis.
Update 10 November 2017:  UKEAT 0056_17_1011: The Appeals Tribunal has confirmed.
Thank you Steve Peers for alerting me to the relevance of the conflict of laws and the Rome I Regulation in particular in the recent Aslam et al v Uber Employment Tribunal decision. The case essentially revolves around whether claimants are employees – it is a pivotal case determining the immediate regulatory context for this part of the ‘sharing economy’. Para 87 is a particularly delightful expression of scepticism towards the sharing economy’s claims (further highlights are here).
Conflict of laws is addressed at para 103 onwards, a completion of the analysis in case of rejection of the tribunal’s view that the UK company in the Uber group employs claimants, and instead one would have to regard Uber BV (of The Netherlands) as employer. I do not think the tribunal expresses itself entirely clearly on Rome I.
If Uber BV is the employer, reclassification of the contract as one of employment (as opposed to one for the provision of services), makes the choice of law for Dutch law partially inoperable (not, as the tribunal notes at para 105 in fine, replaced with the laws on England and Wales). Next the tribunal (paras 106-109) continues to speak of ’employer’ but reviews application of Article 3 (including the application of Article 3(3)’s ‘purely domestic contracts’. If there is a contract of employment, in my view only Article 3(1) and (2) can have any impact on the analysis: the remainder of Article 3 concerns provisions for which Article 8 itself provides exhaustive rules.
From para 110 onwards, the tribunal does more tidily address Article 8 Rome I and holds, after reference to counsel view, that if indeed the Dutch BV is the employer (for it does not suggest that the contract would have to be qualified as one of services), Dutch law would largely apply, except for a limited number of provisions of English law by way of mandatory rules. (Reference to Article 21’s ordre public is justifiably rejected).
I am assuming Uber are appealing. Expect the conflicts analysis to return.
(Handbook of) European private international law, Chapter 3, Heading 3.2.5.